Yesterday I published an article called “Starting The Relationship With A Lie.” It argued that too many sales organizations have adopted a strategy I called fraud as operating model, opening prospect relationships with manufactured fictions, knowing they are fictions, because the math of the meeting-booked rate makes the lie pay.

Within twenty-four hours of publishing, a different company did something worse.

This morning, four people forwarded me an email. The email was sent at 3:14 AM. The From line read “Dave Brock connectors@intelligence.com.” The subject was “Join me on Intelligence.com.” The body, signed “Best, Dave,” said:

“I just joined intelligence.com, a professional network that ranks our connections based on how well we actually know them, allowing for more seamless and efficient introductions and network building. The power behind intelligence.com is its agentic AI workflows that analyze your network and builds your profile according to who you have real relationships with. Every profile is a true reflection of that person’s network, from who they know to how well they know them. Come join me and let’s see what introductions we might be able to make for each other. Best, Dave.”

I did not write that email. I did not authorize it. I did not, in any meaningful sense, “just join” anything. The email was sent, in my name, to people in some inferred contact that intelligence.com, operated by a company called Collective[i], had built by harvesting email addresses I had not used in years.

The four copies forwarded to me went to people who used to work at my company and no longer do. I have no idea how many other recipients exist, addressed to people I do know, who now believe I personally invited them to a product I have never used.

The Lumen email I wrote about yesterday lied to me. This Intelligence.com email lies as me.

That distinction matters, and the industry needs to slow down and look at it, because it is the next stage of what I described yesterday, and it is qualitatively worse on every dimension.

When the Lumen template arrived in my inbox with a fictitious opening sentence about an account I never had, the harm was contained to one mailbox. I deleted the email, the relationship ended at hello, and the cost was mine alone to bear.

Intelligence.com and Collective[i] did not lie to me. Intelligence.com used my name to lie to my network. Every recipient of that email now believes I joined a product, endorsed it, and personally invited them to it.

The damage is not in my inbox. It is in the inboxes of people Intelligence.com sent the emails. It is in the dozens of relationships I have spent decades building, every one of which has just received, from “me,” a piece of marketing copy I would never have written, vouching for a product I do not use, in a voice that is supposed to be mine.

I want to be precise about what is being done here, because the company itself is precise about it. Read the email’s own description of its product: “agentic AI workflows that analyze your network and builds your profile.” That is not me describing what they did. It is them describing it.

They are telling you, in their own marketing copy, that the product’s value proposition is an autonomous system that examines a contact graph the user has not curated, generates a profile the user has not written, and sends invitations the user has not authorized.

The impersonation is not a bug or a hack or an edge case. It is the marketed feature. It is what they sell.

And this is where the AI productivity story everyone in our industry has been celebrating runs into the version no one wants to talk about.

For two years, every sales and marketing leader has been pitched the same dream: AI will do the work of ten SDRs at one tenth the cost, generating personalized outreach at scale, building relationships your team could never build manually.

The pitch is always framed in the language of productivity. The unspoken version is darker. AI does not improve the integrity of outreach. It removes the human cost of the dishonest version.

The lie that used to require an SDR willing to send it can now be sent at zero marginal cost by a workflow that has no conscience, no career, and no professional shame to compromise. The fraud-as-operating-model I described yesterday was bounded by the cost of labor. Agentic AI removes the bound.

Collective[i] is not the only company doing this, just as Lumen is not the only company sending fraudulent prospecting emails. It is representative of a pattern that will get rapidly worse, and the pattern is this: every product you have ever registered for can now impersonate you.

The CRM you tried for a week three years ago. The conference platform you signed up for to attend one webinar. The “networking” tool a colleague invited you to and you accepted because it was easier than declining.

Each of those signups is a stored set of permissions, sometimes buried in a TOS you never read, sometimes simply assumed by a future product pivot, that a vendor can activate the moment they release an “agentic” feature that monetizes your dormant relationships.

You did not opt into being the marketing channel for a product you do not use. You did not opt into vouching for a vendor whose name you barely remember. You did not opt into having your professional reputation extended, without your knowledge, to whatever venture the company you registered with decides to ship next. But that is what is now happening, in your name, to your network, by software you have no visibility into, on a schedule you do not set.

The Lumen leaders I challenged yesterday at least had the modesty to send their lies in their own name. Intelligence.com sends them in mine.

And the people who designed this product, who chose, deliberately, to ship a feature whose core mechanism is the unauthorized use of their users’ identities to deceive their users’ networks, made that choice the same way Lumen’s leaders made theirs: with full knowledge of what the system does, with full understanding of the trust it consumes, and with the rational calculation that the growth math works because the cost is borne by other people.

Which brings me to the question I asked yesterday, and which I will keep asking until the leaders running these operations have to answer it.

What does it tell your customers, your employees, your competitors, and your shareholders that the central capability of your product is the unauthorized use of your users’ names to send communications they did not write to people they did not address?

Not a feature that occasionally misfires. Not an edge case in your AI behavior. The thing you sell. The thing your investor deck describes as your moat. The thing your CEO talks about in interviews as agentic intelligence.

That question has no good answers either.

Yesterday I argued that the honest seller pays the tax that fraud-as-operating-model operators created. Today I have to add a second tax. Every honest seller, every person in this profession trying to build a real relationship with a real prospect, is now also competing against AI systems that will send invitations, endorsements, and “warm introductions” in the names of people who never agreed to give them.

The trust commons I described yesterday is not being depleted. It is being stolen from each of us. The looting is happening at machine speed. And the companies doing it are calling it innovation.

I have requested Intelligence.com’s owners for a complete list of every person they emailed in my name, and the addresses they used. I will write apologies to all of them. That is the cost they have imposed on me, and there is no version of “deleting my account” that undoes it.

Once the lie has been sent in your name, the cost is permanent. The relationship damage is permanent. The cynicism the recipient now carries about every future communication from you, “Was that one really from him, or was that another product doing it again,” is permanent.

That is the world this operating model is building. Every leader shipping a product like this should be required to say, out loud, in front of their customers and their employees and their shareholders, that this is the world they have chosen to build.

Most of them won’t.

Which is why we have to say it for them.

Afterword: For the curious, here one of the emails they sent in my name.

Afterword: Even AI gets pissed off with this behavior. This is a fascinating AI generated discussion of this post.

The email arrived this morning. The opening is worth reading slowly:

“I coordinate meetings for your Lumen account team. I was asked to put 30 minutes on your calendar in the next few weeks.”

I have no account with Lumen. I have never had an account with Lumen. There is no account team. No one was asked to put thirty minutes on my calendar, because no one at Lumen has any idea who I am or what I do.

The first two sentences are, in their entirety, a fiction. They are not careless phrasing. They are not a generic template that happens to read awkwardly. They are a precisely engineered piece of writing designed to manufacture, in twenty-five words, the impression of a relationship that does not exist, an internal team that has not had a conversation about me, and a directive from above that makes the meeting feel pre-decided rather than requested.

I want to be clear about what this article is not. It is not a complaint about an SDR. The SDR did not write this email. Sending these outreaches is the job he was hired to do, it’s what too many are hired to do.

This article is, also, not about Lumen. But using Lumen as an example vividly illustrates the operating strategies that have become too pervasive. Not just start-ups or mid-sized corporations, but large enterprises. You will have had similar experiences with any number of organizations.

But let’s look at what this seemingly innocuous communication really means.

Somewhere in the Lumen org chart sits a person, probably with a title of Sales Development or Demand Generation, whose job is to generate these templates for all the SDRs. They may have run it through some sort of A/B testing, but I doubt it. Their job is to generate outreach, and testing just slows them down. After all, a failing campaign doesn’t mean much, they can just move on to the next campaign.

These demand gen specialists, like the SDRs, have their jobs to do. They have managers supervising their work. Those managers have a VP, and that VP probably reports to someone in a CMO or CRO role.

These managers focus on reviewing and approving the activities. When the math of the results doesn’t work, they move on to the next, crossing their fingers that it will work.

None of these people care that this email landed in my inbox. They know nothing about me and don’t care about me. I’m just one of thousands of people receiving the same email. These people designed the email, and it is performing as expected.

This is what every GTM leader needs to face head-on. The lie in the opening sentence is not a defect of execution. It is the strategy too many have adopted. Lumen, in this case, has decided that defrauding strangers at the top of the funnel produces enough qualified meetings to justify the customers it loses, the reputation it burns, and the trust it consumes.

The leaders who approved this template did not stumble into it. They may not have even tested it. They watched the meetings get booked, didn’t ask how, and created more outreaches like it. They made the rational economic decision to mislead the targets of their outreach.

There is a name for that decision, though no one will use it. It is fraud as operating model. Fraud may seem a strong word. We usually associate it with criminal activities. But fraud in this outreach is entirely appropriate and the word has been earned. This outreach, by design, is intended to mislead the recipient.

I name Lumen here because the email is real, the company is real. This is an operating strategy they have chosen, and they must own the consequences of that strategy.

And consider the consequences they are choosing to own. Lumen’s stated values, published on the company’s own materials, include “Teamwork, Trust, Transparency” and “Customer Obsession.” Its CEO, Kate Johnson, has said publicly: “At Lumen, we are putting our customers at the center of everything we do, recognizing they want and deserve fundamentally better experiences from telecom companies.”

These are not aspirations the company is privately falling short of. They are the words the company chose to describe itself in the same period it deployed the email I received this morning.

The gap between the language and the operating model is accidental. It is a design. The values exist to be communicated. The fraud exists to be executed. They coexist because leadership decided they can. The customers Kate Johnson says are at the center of everything are not, in fact, anywhere near the center.

The center is occupied by a meetings-booked metric, and the customers, actual or invented, like me, are inputs to that metric. The PR language and the prospecting template are two outputs of the same calculation: “What can we say that sells, what can we send that books meetings?” Truth is not a variable in either equation.

But Lumen is not unusual. It is representative of what we see too many organizations adopting as their prospecting and customer strategies. Everyone reading this article can produce, from their own inbox, three or four examples sent within the last week that operate on identical principles.

The fake-familiarity opener. The invented internal context. The “as we discussed” when nothing was ever discussed. The “per our conversation” when no conversation occurred. The “I was asked to reach out” when no one asked anything. The “your team mentioned” when there is no team and they mentioned nothing.

These are not market variations. They are not company-specific strategies. They are the same template, rewritten by thousands of organizations because each of them is optimizing the same metric against the same operating model.

This model rewards manufactured intimacy at scale.

It’s important to call this out, describing it plainly, because the people who run this operating model rarely do. An SDR is hired, given a quota for meetings-booked-per-week, and provided with a sequence of templates engineered for response rate rather than truth.

The templates are written by marketing, optimized by ops, and approved by sales leadership. The SDR is measured on volume and conversion, not on the integrity of the opening sentence. The SDR’s manager is measured on the team’s meeting-booked rate, which feeds the AE pipeline, which feeds the CRO’s quarterly forecast, which feeds the board’s revenue narrative.

At no point in this chain does anyone get measured, compensated, or recognized on whether the first interaction with the customer was true. Truth is not in the comp plan. Meetings booked is in the comp plan. The system produces exactly what it pays for.

What the system does not measure, and what no leader running this playbook cares to measure, is the cost. Not the cost of developing and running the programs. Not even the cost of annoyed prospects deleting these emails.

The cost is the systematic destruction of the trust critical in building relationships in the target markets.

Think about the prospecting outreaches we care about and want to respond to. Consider the seller you would actually want to hear from. The one who understands your company, has done the research, who has taken the time to learn about you. The one that has a genuine reason to believe she can create value for you. The seller who wrote her own email, who opened with something specific to your business, who proposed a conversation rather than presumed one.

Those sellers exist. She is your competitor’s best rep, or the specialist who solved exactly the problem you are wrestling with right now. Her problem is not that she has done too little work. Her problem is that her email arrives in the same inbox, on the same morning, as Lumen’s SDR’s. And it is automatically routed to the same spam folder you created months ago, after the hundredth email convinced you that no cold outreach was worth reading.

That is the real damage. The fraud-as-policy operators just failed to book meetings. They are training every recipient of these outreaches to spam 100% of them. They have made all of us stop reading. They have made all of us stop replying.

They have made the cost of reading cold outreaches higher than the expected value of doing so. As a result, the careful, well-researched outreach is now handled in the same way as the fraudulent kind. Each of these “poison the well” for everyone else.

The honest seller is paying the penalty that Lumen and its peers have created. The honest seller did not contribute to this, but they cannot opt out of it. The honest seller’s career, and our own ability to access valuable information, are degraded by all those flooding the markets with meaningless, fraudulent outreach. They don’t recognize the damage they do both to themselves and everyone trying to participate in a meaningful way.

This is what makes the operating model corrosive. Organizations running a fraud-as-policy outbound machine at enterprise scale harms every other seller in the market, and every other buyer in the market, because the trust he destroys. These organizations draw down a shared belief takes decades to build.

There is a second cost to the offending company, that the operating model ignores. The lie at “hello” does not stop there. Once a buyer suspects that the opening sentence of the relationship was manufactured, every subsequent interaction is judged based on that initial experience. “Are they continuing to mislead me? Can I believe their offer if their first sentence was a manipulation? If I’m currently using the products, can I believe what they are saying about the renewal?”

The lie at the beginning of the conversation does not vanish when the deal closes. It is embedded into the relationship, shaping it over time. The CRO who optimized for meetings booked has, in the same motion, optimized against customer lifetime value.

And there is a third cost, which is the one the leaders should care about most and rarely do. The SDRs themselves know. SDRs have the same brain their prospects have. They understand their first sentences are not true. They rationalize this as just part of the job, they hit their metrics and comply with what leaders are telling them to do in their engagement strategies.

These SDRs learn what their companies have taught them. They have learned the way you open a customer relationship is with a manufactured fiction directed by the demand gen experts and supported by the leaders.

This is what leaders are training the next generation of sellers and future leaders to do. We shouldn’t be surprised as entry level SDRs are moved into AE roles and then into management, that they continue to practice what they learned in developing these fraudulent relationships.

It is embedded into our GTM strategies and practices, it is taught to everyone in the organization, it becomes the modus operandi. It becomes the company culture.

Which brings me back to the question I would put to the people at Lumen who designed and approved the email this SDR sent me this morning, and to every sales leader running an equivalent operation.

“What does it tell your customers, your employees, your competitors, and your shareholders that the first sentence you chose to send to a stranger was a sentence you knew was not true? Not a sentence that turned out to be wrong. Not a sentence that was generated by a system you don’t fully control. A sentence you wrote, you tested, you optimized, you approved, and you deployed, knowing precisely what it said and precisely what it implied and precisely how false both were.”

That question has no good answers. The leaders running these operations know it has no good answers, which is why the question is never asked.

This will not change just because someone like me is writing about it. All the strategies, plans, metrics, comp structures are built to reinforce this fraudulent outreach. Everything that rewards this will make “fraud as a policy” the GTM strategy for those organizations.

What may change this is buyers. The people that write checks to us. We see this every day and it is escalating. They don’t want to see us or hear from us. They don’t trust what we say, when they do meet us. They are slower to commit, much more cautious, much more skeptical.

It costs everyone, buyers, sellers, and the markets in which we compete. And it takes years to recover what was lost.

The first sentence, the first outreach is so important. It shapes every other communication we have with our prospects. It shapes our customer and our futures. Until we recognize this, the costs continue to grow.

And until buyers make that cost legible, my inbox will continue to fill with messages like the one Lumen’s SDR sent this morning. And every honest seller trying to reach me, and every honest seller trying to reach you, will continue paying the tax that Lumen and its peers have decided is someone else’s problem.

Afterword: This AI generated discussion is one of those that brings much greater clarity to these ideas, than the article itself. It’s worth listening to. Enjoy!

We’ve been through decades of pricing experiments. Enterprise licenses sold access to software that mostly sat unused. Seat-based pricing rewarded headcount, and shelfware and bloated tech stacks became the running joke of the SaaS era. Consumption-based pricing tied cost to activity, better until buyers realized activity is not value.

Now we’ve arrived at what the industry calls value-based or outcome-based pricing. The vendor commits to a unit of business value; a resolved ticket, a qualified lead, a closed deal, an hour reclaimed, and prices against that unit. The pitch is clean. The vendor only wins if you win. If resolved tickets are the desired outcomes, payment is based on the number of resolved tickets.

Shelfware disappears. The CFO finally gets a line item that ties spend to result.

This is not hype. Salesforce Agentforce reportedly reached around $800 million in ARR running three pricing models at once. Intercom Fin charges $0.99 per resolved conversation. HubSpot moved Breeze to $0.50 per resolved conversation in April. Adobe announced outcome-based pricing for its CX Enterprise AI suite.

Bessemer’s tracking across two hundred AI vendors shows hybrid pricing (a base fee plus outcome or usage) went from 27% to 41% adoption in twelve months. Seat based pricing fell from 21% to 15%. Gartner projects 40% of enterprise SaaS will include some outcome-based element by year-end, up from 15% two years ago.

The movement is real. The question it raises is one nobody is yet willing to answer.

Who gets credit?

Picture a salesperson closing a deal. She uses five tools to do it. The CRM holds the account history. The sales engagement platform sequences her outreach. Conversation intelligence coaches her during calls. The forecasting tool flagged the deal as winnable last quarter. CPQ generates the proposal.

The deal closes. Revenue lands.

At renewal, every one of those five vendors arrives with a slide. Each shows how their tool contributed to the result. Each has data. Each has a number. CRM claims credit because of the account history. The sales engagement platform claims credit for sequencing the outreach. Add all the claims across a typical enterprise tech stack and you get two or three times the actual revenue. Everyone is claiming credit for the same deal.

This is the math problem value-based pricing creates the moment you have more than one vendor in the stack, which is to say, immediately. The contracts each assume the vendor caused the outcome. The outcomes are not separable. The buyer is left holding five invoices that, taken together, claim to have produced more value than the business actually generated.

We’ve seen this before.

The B2B marketing attribution category spent twenty years trying to answer who deserved credit for a closed deal. First-touch. Last-touch. W-shaped. U-shaped. Weighted multi-touch. Considerable venture capital chased the question. Matt Heinz published the postmortem last week.

The dashboards were defensive theater. Nobody believed the numbers. CMOs deployed the platforms and then made budget decisions on instinct anyway. The premise, that you could pixel-track a buying committee’s three-to-nine-month decision process, never survived the reality of how B2B buying actually works. CFOs eventually stopped writing the checks, and the category quietly collapsed.

Value-based vendor pricing is the same problem moved up one layer. Instead of nine marketing touchpoints arguing on a dashboard, you have five or six vendors arguing on a renewal slide.

Instead of one CMO defending a budget to one CFO, you have an organization defending a stack of contracts where every vendor has data showing they were the cause. The attribution problem didn’t get easier going up a layer. It got harder. Now the attribution claim is also the invoice.

Definitions drift toward whatever is easiest for the vendor to claim. The question of whether the outcome would have happened anyway never gets asked, because neither side benefits from asking it. The vendor with the best tracking data, not the vendor producing the most value, writes the biggest invoice.

That’s the first level of the credit problem. Vendor versus vendor, each claiming the same closed deal.

Now add the people who actually did the work.

The AE built the relationship over eighteen months. The SDR set the meeting that started it. The manager coached the rep through the hard moments. The product team built what the customer bought. The CEO signed the partnership that made the prospect take the call in the first place. Every one of those claims is partly true.

None is completely true. Sales compensation plans exist because organizations gave up on a clean answer and picked a metric the comp plan can pay against. The metric isn’t the truth. It’s the workable approximation everyone has agreed to live with.

Value-based vendor pricing imports this fight into pricing strategies.

If Agentforce gets paid for the qualified lead, what does the SDR get paid for? If Fin resolves the ticket, does the support rep who wrote the knowledge article Fin learned from get credit for the resolution? If the coaching tool gets paid for the deal it influenced, does the manager who coached the rep for three years before the tool showed up get nothing?

Procurement is signing contracts whose attribution logic competes directly with the compensation plans and metrics Rev Ops is administering. The two functions are not talking to each other about it. The vendor’s claim and the employee’s claim are now drawing on the same pool of credit, and the organization has not decided who is right.

That’s the second level. Tools versus people, each producing the same outcome. Each claiming credit for that outcome.

This is where the arrangement starts to break.

The procurement leader who signed the value-based contract has to defend the spend to the CFO. The CFO has to defend the line item to the board. The story they tell, “we paid $X and got $Y in attributed pipeline, here’s the ROI slide,” relies on the same instrumentation that broke at the campaign level in marketing attribution. The buyer has a vendor dashboard saying outcomes were delivered. No independent way to verify it. No way to answer whether it would have happened anyway.

Right now the bills are small enough that nobody is asking. That changes fast. AI-native application spend at large enterprises jumped 393% year over year. When the invoices get large enough to matter, the same CFO who killed the attribution category is going to ask the same question of the value-pricing category. What did we actually get for this?

The answer will look familiar. The dashboard says one thing. The people closest to the work say another. Nobody can tell which is true.

Gainshare consulting tried this in the 1990s. Performance-based advertising tried it in the 2000s. Most of those arrangements collapsed back to fee-for-service. This collapse was not because it was wrong, but because the attribution math underneath never held up to under closer investigation. These collapses took years. Attribution arguments always take years to play out. You need enough renewal cycles and enough invoice inspection for the math to fail.

The harder reality is the one we keep avoiding.

Outcomes in complex selling are not produced by tools. They are produced by people using tools, inside organizations that have made years of decisions about culture, hiring, training, coaching, and relationship investment.

The vendor claiming credit for a closed deal is doing what a bad manager does, taking credit for the team’s work. The buyer who accepts that framing is doing what bad executives do, outsourcing accountability to a vendor.

Value-based pricing only makes sense if you believe the tool is the actor, that it has no dependency on the people or the other tools around it. That is the only scenario in which the contract holds together.

It is not the only situation in which the business holds together.

Capable people produce outcomes with help from tools. But vendors pretending otherwise will keep claiming credit. The dashboards will keep claiming credit. Somewhere down the line, a CFO is going to ask what we actually bought. The answer will look a lot like the answer marketing attribution gave when its moment came. Less than we paid for last year. Probably less still next year.

The pricing model that finally works will be the one that stops pretending it can answer a question nobody has ever answered. Until then, everyone in the system has an incentive to claim credit for the same dollar. And value-based pricing institutionalizes that incentive.

Afterword: This is a stunning AI based discussion of this post. I love their different take on my article. They make the key point of “Who deserves credit, “ in a very powerful way.

Two things have been true in most organizations for a long time, and both are now coming due at once.

The first is that the people running these organizations don’t have an accurate picture of what’s happening in them. Leaders consistently believe their staff are more aligned with business goals than employees report being. Leaders believe their communications are clearer than employees experience them as. Leaders believe their teams understand the strategy. Most employees say they don’t. Half of employees admit they’re giving only the bare minimum. This has been documented in survey after survey, year after year, and it doesn’t move much.

The second is that an entire layer of management infrastructure was built over the last twenty years focused on letting leaders feel in control without actually being close to the work. Metrics replaced judgment. Dashboards replaced understanding. Activity tracking replaced apprenticeship. Compliance frameworks replaced craft. Each one was sold as professionalization. Each one in aggregate moved leaders further from the work their organizations were actually doing, while giving them more confident-looking reports about it.

Those two things produced a generation of leaders running organizations they couldn’t actually see. The activity layer kept producing dashboards that looked reasonable. The teams kept producing revenue that justified the dashboards. Quarterly earnings kept showing up. Boards stayed satisfied. Nobody was lying. Everybody was responding rationally to a system that rewarded the appearance of insight more than the substance of it.

Then AI arrived, and the same leaders who had lost touch with their organizations were suddenly making large investment decisions about how to deploy it.

This is the cause behind the pattern I described in the first two pieces (The AI-Native Illusion and We Don’t Start With AI). It’s not that leaders bought Interface layers when they should have bought intelligence. It’s not that they failed to ask what their business was for. Both of those are true, but they’re symptoms.

The deeper cause is that leaders were making AI deployment decisions from a fundamentally inaccurate picture of their own organizations. They didn’t know what the work actually was. They didn’t know what their people were actually doing well or poorly. They didn’t know where excellence lived in the organization or what conditions produced it.

So they deployed AI against the picture they had, which was the picture the dashboards gave them, which was the picture the activity layer was designed to produce.

You can’t make good decisions about AI deployment from that activity base orientation. As we’ve seen, this only allows you to optimize on the activities, but they don’t tell you if they are the most meaningful activities. And this is what the majority of companies have been doing.

The dashboards look better. The activity numbers go up. The bills go up faster. The customer outcomes don’t move. The 0.4 to 0.8 percent productivity gains, the 55 percent value realization, the 94 percent of companies seeing no significant value. These aren’t AI failures. They’re the predictable result of deploying expensive new capabilities against a model of the work that was never accurate.

This is also why the standard responses to the disappointing results don’t work. “Be more ambitious” doesn’t work because more aggressive deployment against a wrong GTM model just produces the wrong outcomes faster.

“Eliminate the activity layer” doesn’t work by itself because the activity layer was a symptom, not the cause; remove it and you still have leaders who can’t see the work, and now you have nobody producing the dashboards either.

“Bring in better tools” doesn’t work because the problem was never the tools. The problem was that nobody at the top knew what excellence in the work looked like, and you can’t deploy tools against a standard you don’t have.

The honest version of this is uncomfortable. A lot of senior leaders running companies right now built their careers in the activity layer. They became excellent at managing dashboards, running operating reviews, hitting quarterly numbers, producing the kind of organizational legibility that boards and investors reward.

They didn’t become excellent at watching the actual work, developing the people who do it, or understanding what produces customer outcomes that matter. The system they came up through didn’t require those capabilities. The system they’re now being asked to lead does. Most aren’t equipped for it, and the AI deployment decisions they’re making reflect that gap.

This isn’t an issue of intelligence. It’s the result of the way we have seen leaders shift their focus over more than two decades. These leaders who built the activity layer were responding to the incentives of their time.

Those incentives have now changed. The pressure that was implicit before AI is explicit after. Leaders who can’t see the work clearly are going to make worse and worse decisions as the cost of those decisions accelerates. The ones who can see the work, or who do the work to learn how to see it again, are going to make better ones.

That’s the cause. Now the alternative.

The first move is the hardest, and it has nothing to do with AI. Leaders have to rebuild their connection to the actual work their organizations do. Not symbolically. Not through better dashboards. But through actually being in the work, watching it happen, doing pieces of it themselves, talking with the people who do it every day about what makes it hard and what makes it good.

This sounds like 1990s management consulting and it gets dismissed for that reason. It shouldn’t be. It’s the prerequisite for every other decision a leader is now being asked to make.

A CRO who hasn’t sat through real customer discovery calls in two years can’t tell you what excellence in discovery looks like. A CEO who only sees customer conversations through quarterly NPS scores can’t tell you where customer relationships are actually being built or lost. A CHRO who experiences her own organization through engagement survey rollups can’t tell you which managers are developing people and which are running compliance theater.

None of these leaders are stupid. They’re just operating from a model that rewarded focus on the activity layer. This model was designed to make management feel in control without requiring contact with the work.

The first thing leaders must change is how they engage in and understand the real work that needs to be done. Block calendar time for it. Sit in real customer conversations as an observer. Ride along with frontline workers. Read raw customer feedback, not summaries. Listen to actual sales calls, not curated highlights. Watch a real deal review, not a dashboard. Do this regularly, until you can describe in operational terms what excellence in the core work of your organization looks like, what good and bad versions of that work feel like in practice, and where the conditions that produce excellence currently exist or have been hollowed out.

Until you can do this, every AI investment decision you make is going to be a guess, and based on what we’ve seen so far, it is likely to be wrong!

The second move is to revisit the foundational questions from the second piece in this series: what is our business actually for, what customer outcomes do we exist to create, what excellence does that require, what conditions produce that excellence. But now you revisit them equipped with what you’ve learned from being in the work.

Let me pause for a moment. In looking at these foundational questions, we tend to answer them based on our internal corporate goals. For example, we tend to answer the question, “what is our business actually for,” in financial terms like revenue goals, profitability, and growth. These are only the outcomes of the real reason our businesses exist and the customer outcomes we create that make them choose us. Fall into the revenue targets, profitability and growth traps and you are still focused on the wrong things.

When we spend time understanding the work, these foundational questions stop being abstract. You can answer them in concrete terms because you’ve seen what your organization actually does, where it creates value for customers, where it doesn’t, what your people are capable of, where the apprenticeship still works, and where it’s been replaced by activity that produces nothing.

Here’s where the argument for AI gets sharper than I’ve been able to make it before. AI isn’t just something to deploy after you’ve answered the foundational questions. AI is the reason you can answer them differently than you could have ten years ago.

The constraints that shaped what businesses thought they could do, what they could build, what they could deliver, who they could serve, at what cost, at what speed were all based on assumptions about human capacity in specific roles.

Those constraints have shifted. Some have disappeared entirely. The customer outcomes you couldn’t reach before because the unit economics didn’t work, you might be able to reach now. The work you couldn’t do because you couldn’t hire enough specialists, you might be able to do now. The relationships you couldn’t sustain at scale because the human time wasn’t available, you might be able to sustain now. The question “what is our business for” has answers in 2026 that didn’t exist in 2020, because the things a business can be for have expanded.

And we’ve experienced this before. The internet/web completely shifted how much of work was done. The early IT revolution, as well as the industrial revolution allowed us to reimagine businesses, offerings, redefine our markets/customers, transform our organizations and work. They allowed us to achieve things we never knew possible.

With AI, we are going through the same transformation we’ve experienced before, the technology is vastly different and the timeframe to make the changes can be dramatically compressed.

This is the part that the efficiency framing misses entirely. Efficiency thinking takes the existing business as given and asks how AI makes it cheaper to run. The thinking shift takes the existing business as a starting point and asks what the business could become now that AI has changed what’s possible. Those produce different roadmaps. Most companies are running the first one. The few that are running the second are going to look in five years the way Amazon looked in 2003.

The third move is to make the rethinking iterative and experimental, because it can’t happen any other way. You can’t sit in a strategy off-site and figure out what your business should be in an AI-shaped world. The capability set is moving too fast. The customer expectations are moving too fast. Your own understanding of what’s possible is going to develop only through actually trying things and seeing what they reveal.

The right model isn’t “rethink the business, then deploy AI.” It’s “form a working hypothesis about what the business could be, deploy AI to test the hypothesis, learn what’s actually possible, refine the hypothesis, deploy again.” The rethinking happens in motion. The deployments are how you learn.

This means leaders need to do something most haven’t done in years: sponsor real experiments with real budgets and real expected learning, not pilots designed to produce a case study for the board.

Pick a customer outcome that matters. Form a small team that includes both senior leaders and frontline people with knowledge of the actual work. Give them real authority to deploy AI against that outcome in ways that go beyond automating what already exists.

Set the success measure as customer outcome change, not adoption rate or productivity proxy. Run it for a defined period. Read the results carefully. Adjust the hypothesis. Run it again.

Most companies don’t do this because the activity layer’s incentives reward predictability over learning, and because most leaders don’t have the in-the-work knowledge required to design the experiment well. Both of those obstacles are exactly what the first two moves are designed to address.

The fourth move is to be ruthless about distinguishing between work that produces customer outcomes and work that exists because the activity layer needed it to exist.

AI makes this distinction enforceable in a way it never was before. When you can see what work actually produces customer outcomes, you can let AI take the work that doesn’t, while preserving and protecting the work that does.

The pipeline review where a manager challenges a rep’s thinking, preserve. The pipeline review where someone reads a dashboard out loud for an hour, eliminate. The customer conversation where a rep develops the relationship knowledge that produces the next deal, preserve. The CRM data entry that exists because a manager wants to feel in control, eliminate. The deal post-mortem where the team learns from a loss, preserve. The forecast where a rep manufactures numbers that match what her manager wants to hear, eliminate. The distinction is operational, not philosophical, and you can only make it from inside the work.

None of this is new, we have seen this in consistent high performing organizations for decades. But without this, we can’t begin to imagine the true power of how AI can help us redefine our business.

The fifth move, and the one I’ll take up directly in the next piece, is to figure out what the new apprenticeship looks like. Some of the work that built judgment historically is going to be done by AI now, and that’s appropriate for some of it and disastrous for the rest.

The leaders who can tell the difference, and who deliberately preserve and rebuild the conditions that develop the next generation of capability, will have organizations capable of excellence in 2030. The leaders who don’t will have organizations that look efficient on paper and have no internal capacity to produce anything new.

None of this is fast. None of it has a vendor answer. None of it is what the board is going to reward in the next earnings call. It’s also the only path that produces a business worth running five years from now.

The leaders who started this work last year are already pulling ahead. The leaders who start it this year will catch up. The leaders who keep doing what they’ve been doing, running the activity layer, deploying AI against bad pictures of their own organizations, optimizing dashboards while customer outcomes drift, are going to spend the rest of the decade explaining to their boards what happened to companies that looked solid on paper but failed in the real world.

The cause was always the same. Leaders lost contact with the work. The system was designed to let them.

AI is now spending real money to make the consequences of that disconnection impossible to hide. The fix isn’t more AI. The fix is rebuilding the connection between leadership and the actual work, asking the foundational questions from inside that connection, running real experiments to learn what’s now possible, and using AI as the test instrument rather than the destination.

That’s the work this moment is asking for. Almost nobody is doing it. The ones who do will run the businesses that the rest of us write case studies about.

Afterword: A fascinating AI based discussion of this post. I like the different analogies they use to help reinforce the points in this article. Enjoy!

Companies are spending more on AI than ever before, even as the cost of AI itself has fallen. The price of running an AI query dropped 80 to 90% over the last two years. Total enterprise AI bills are skyrocketing anyway. Average enterprise AI spend rose from $1.2 million in 2024 to $7 million in 2026. The cost of running AI in production now consumes 85% of the AI budget at companies operating at scale.

Agentic workflows, the ones being marketed as transformation, consume 5 to 30 times more compute per task than the chatbot pilots that came before them. Earlier this year, Uber’s CTO told The Information that the budget he thought he would need was already blown away.

Across all of this, value realization sits at 55%. McKinsey’s 2025 survey found 94% of companies seeing no significant value from AI investment. Productivity gains where they exist range from 0.4 to 0.8%.

The standard explanation is that companies aren’t deploying aggressively enough. The argument runs: if you’re seeing marginal returns, it’s because you bought interface layers when you should have bought intelligence. You ran pilots when you should have eliminated stacks. You added AI to existing processes when you should have rebuilt the organization around what AI makes possible.

Be more ambitious. Move faster. Replace, don’t augment. Let the recursive engine compound.

Parts of that argument are right. Companies that build adaptive systems do pull ahead, and the gap widens over time. Adding AI to existing processes is exactly the failure mode I described in The AI-Native Illusion. Replicating poorly thought-out processes at scale will produce mediocre to poor results at scale.

But the high-ambition version of the argument has the same omission as the Interface-native version it claims to surpass. Both are answers to the question, “what AI should we deploy.” But these questions miss something. We are missing the foundational questions that are part of every transformation. We must address these first. They are:

What is our business actually for? What can we now do that we never could do before? How is it important to our customers, our markets, our organization?

Those sound like questions any senior leader can answer. They aren’t. Most cannot answer them without retreating to the standard corporate or investor decks. They talk about mission statements or product descriptions or revenue targets. The mission statement is not what the business is for. It’s what the business says it’s for, in the language that survived legal and PR review. Revenue is not what the business is for either. It’s the result of doing something useful for someone who’s willing to pay. Product descriptions describe what we sell, not what we’re for.

What the business is for is the answer to a harder question: what customer outcomes do we exist to create that wouldn’t exist without us? Not what do we sell, not what do we measure, not what’s on the website. What changes for the customer because we exist, and out of all the things we could do to produce that change, what are we distinctively positioned to do better than anyone else?

Underlying this, is what do we want to stand for with our customers, employees, investors, and community?

Companies that can answer these questions can have useful conversations about AI. Instead, everyone skips over these, starting in the middle of a discussion without considering these prerequisites.

Part of the reason this is done is the fundamental questions are tough work. They involve deep understanding, thought, and discussions in the organization and with customers. Part of the problem is momentum. We focus tactically on what’s next and how we do more, seldom revisiting these foundational questions.

This is where the cost data starts to make a different kind of sense. The exploding bills, the marginal productivity returns, the value realization at 55% are not symptoms of insufficient ambition. They are symptoms of organizations deploying AI without having answered the foundational question.

When you don’t know what your business is for, you can’t tell which AI deployments serve it. You deploy what’s available, what’s marketed, what the vendors are pitching, what looks like motion to the board. You optimize for measures that focus on activity rather than outcomes. The bill goes up because nothing in the deployment decision is calibrated against actual purpose.

The high-ambition playbook makes this visible in a particular way. Read the typical case study. The story is usually that AI eliminates the “low-value” work, logging, prep, pipeline review, follow-up drafting, contact research, forecast assembly, so people can spend more time on the high-value work, usually defined as customer interaction. The metric that gets celebrated is something like, “sellers went from 30% selling time to 80% selling time.”

That metric is the tell.

Selling time is not what sellers produce. Selling time is an input, badly measured. The output of a seller is customer outcomes. Relationships built. Problems solved. Deals that close because they should. Deals that don’t close because they shouldn’t. Accounts that grow because the seller understood the customer’s business well enough to find the next thing that mattered.

What’s also missed in what’s called “low-value” work is the knowledge and judgment that’s built in doing it. How do we effectively communicate with and connect with our prospects and customers? How do we most effectively leverage the research in our meetings? What are the factors that impact the forecast of a particular deal? The numbers and the data don’t contain the experience, judgment, and insight. AI can certainly contribute, but the idea that all of this is low-value demonstrates a huge misunderstanding of the work itself.

When you optimize for selling time, you have measured the wrong thing. You’ve assumed that more time in front of customers automatically produces more value created with customers, which is only true if the seller has the judgment to make that time count.

The judgment is the part that doesn’t get discussed.

Sales, and this generalizes to most knowledge work, isn’t a set of discrete tasks. It’s an apprenticeship in judgment. The pipeline review where a rep gets challenged on a deal by a manager who has seen a thousand deals, that’s where the rep learns to read a deal. The prep work where a rep researches an account before a call, that’s where she develops the business and relationship knowledge that lets her hear what the customer is actually saying. The follow-up the rep drafts herself, that’s where she learns what trust feels like in writing, and what the cost of getting it wrong looks like when she doesn’t connect in a meaningful way. The forecasting process where she has to defend her numbers, that’s where she develops the discipline to know what she actually believes versus what she’s hoping for or the report is telling her to do.

Strip those tasks out and the activities are gone, but so is the apprenticeship. The seller who never builds her own pipeline view never learns to read one. The seller whose customer research is automatically given to them never develops the relationship knowledge that makes intelligence useful in front of a customer. The seller whose forecast is assembled by AI never develops the discipline that produces forecasts that matter.

Keenan made a version of this point recently, quoting a head of enablement who said the quiet part out loud: we trained reps on tools for years, we never really trained them to be good at the conversation, and now the conversation is what we’ll need them to do. He’s right about what’s been happening. The activity layer the SaaS economy built around sellers absorbed the energy that should have gone into developing judgment. AI is now exposing what was actually being produced underneath.

The high-ambition playbook treats all of that lost development as overhead. Cost. Friction. Stuff to be eliminated so the human can do the “real” work. But the real work doesn’t separate cleanly from the apprenticeship. The capacity to do the real work is built in the very tasks being labeled low-value. Eliminate them and you don’t get a seller with more time to be excellent. You get a seller who never developed the capacity to be excellent and now has time on her hands.

And we’ve seen this before AI showed up. Strictly scripting people. Using dashboards without understanding what they mean. The inability to truly understand and engage customers with impact. We’ve watched years of performance decline as a result. AI taking the work away doesn’t improve the skills of the people executing what remains.

This is the people question that the AI conversation systematically avoids. It avoids it because addressing it would force a much harder set of questions. What are we actually trying to develop in the people who work here? What capabilities does the work itself produce that we’d lose if we automated the work? Where does AI extend human capability versus replace the conditions that build it? Where is the apprenticeship in our roles, and what happens when AI removes it?

These questions are uncomfortable because they don’t have vendor answers. They require leadership to know what excellence looks like in the work, and to know where it gets built. Most leadership doesn’t know, not because they’re stupid, but because the conditions that produce excellence have been eroding for years. The metrics-driven, activity-managed, compliance-scripted version of most knowledge roles already removed most of the apprenticeship. AI is finishing what activity tracking and call recording started.

So when we talk about starting with the business rather than starting with AI, this is what we mean. It isn’t a process recommendation. It’s a recognition that the question “what AI should we deploy” only has good answers when we’ve already answered the prior questions. What is the business for? What customer outcomes does it exist to create? What can we do that we’ve never been able to do before? What excellence does that require from the people who do the work? What conditions produce that excellence?

These take real strategic thinking. Most companies don’t do real strategic thinking. They do market positioning, financial planning, and roadmap prioritization, all of which are downstream of strategy but not strategy itself.

When real strategic thinking is happening, the dialogue between strategy and capability becomes natural. You know what you’re trying to be excellent at, so you can evaluate which AI capabilities actually serve that excellence, which substitute for it, and which would erode the conditions that produce it. The deployment becomes a portfolio of choices made against a clear standard.

When real strategic thinking isn’t happening, the dialogue can’t happen because half of it is missing. There’s no clear answer about what the business is for, so there’s no standard against which to evaluate AI deployment. The vendors fill the vacuum. The metrics fill the vacuum. The board’s questions about AI fill the vacuum. The deployment becomes whatever survives the budget cycle and produces the cleanest case study, which is usually whatever maximizes what’s measurable rather than what matters.

This is why the cost data looks the way it does. Seven million dollars in average annual AI spend. Fifty-five percent value realization. Productivity gains under one percent. Bills running ahead of every projection finance teams made.

Organizations aren’t being stupid. They’re responding rationally to incentives in an environment where the foundational strategic question never got asked. The vendors have done well. The boards are satisfied. The bills are coming due.

There’s a version of this where a leadership team actually sits down and answers the questions. What customer outcomes do we exist to create? What excellence in our people produces those outcomes? What conditions develop that excellence? Where does AI amplify what we’re already trying to be, and where does it replace the conditions that produce it? Where is the apprenticeship in our work, and what happens when AI removes it? These aren’t AI questions. They’re business questions. The AI questions only become tractable after they’re answered.

Almost no leadership team is doing this work. Some can’t because they don’t have the capacity for that kind of strategic thinking. Some won’t because the answers would force changes they’re not willing to make. Some are too busy responding to vendor pitches and board pressure to make the time. The result is the same. AI deployment without strategic foundation. Activity without purpose. Bills without returns.

The question worth asking isn’t “what AI should we deploy.” It’s whether anyone in the organization is doing the prior work that gives AI deployment any meaning at all.

If they are, the deployment will pay off, eventually, in ways that show up in customer outcomes rather than in productivity dashboards. If they aren’t, no amount of ambition will save the deployment from the same fate as the Interface-native crowd. Different mechanism, same disease, same bill.

Afterword: This is the AI generated discussion of this post. It’s fascinating, they start with an example of dropping a F1 Car onto a dirt road, saying the car can’t achieve it’s potential because of the limitations of the road. They bridge into a fascinating discussion of this post. Brilliant! Enjoy.

Most companies marketing themselves as AI-native aren’t.

They’re Interface-native. There’s a difference, and the difference matters more than anyone in the current hype cycle wants to admit.

Strip the AI-native label off most of the companies proclaiming this and look at what’s actually happening.

For example, the AI SDR startup that raised at a billion-dollar valuation is orchestrating LinkedIn Sales Navigator for identity, Apollo or ZoomInfo or Clay for enrichment, Outreach or Salesloft for delivery, and writing back to Salesforce for the system of record.

The AI is doing the personalization and the orchestration. But the underlying systems are entirely traditional. Pull any one of those underlying systems out and the AI SDR has nothing to operate on.

The same is true of the AI revenue intelligence platforms ingesting from Salesforce and Gong, the AI sales coaching tools analyzing what Gong already captured, the AI agentic platforms automating workflows between SaaS tools that have existed for fifteen years.

This isn’t a critique of these tools, some are quite useful. The personalization is real. The orchestration is real. The efficiency gains are real. But none of it is transformation, and the “AI-native” marketing by these organization is positioning them as transformation.

We aren’t actually buying AI-native systems, we are buying an interface layer. It is a much slicker, smarter, faster way to access systems that already existed. These tools provide much easier ways to integrate disparate databases and workflows. Using traditional tools, it takes months or years to achieve the same level of integration. So these Interface-native tools can be very powerful.

But these Interface-native tools are not the future of enterprise software, though many of them imply this in their marketing. The reason is, without the current enterprise software platforms, they would crumble. They would cease to exist.

The traditional platform vendors are making similar claims, but from a slightly different perspective. Salesforce shipped Agentforce and has recently announced Headless 360. HubSpot shipped Breeze. ServiceNow has its agents. Microsoft has Copilot threaded through everything. Each was launched with the marketing language of fundamental rethinking.

Architecturally, each is an agent interface on top of the same data model, the same workflow engine, the same business logic that existed before. Salesforce in 2026 with Agentforce is still Salesforce. The object model didn’t change. The fundamental approach to managing customer relationships didn’t change. What changed is how you access and act on what’s already there. Basically, a slicker user interface. There’s value in that. Sometimes substantial value. But it’s interface value, not architectural value.

And the vendors marketing it as transformation are doing the same thing the startups are doing, selling efficiency disguised as something it isn’t, at least yet.

Here’s where it gets interesting for buyers. The startups in the first category, the AI SDRs, the AI revenue platforms, the AI coaching platforms, the agent orchestration companies aren’t just dependent on the underlying platforms. They’re competing with those same platforms.

When Salesforce ships interface capabilities that overlap with what an AI SDR startup, the AI revenue platforms, or the AI coaching tools provide, the startup’s value proposition compresses fast. The startup has no defense because its entire architecture depends on Salesforce, and Salesforce can offer the same interface layer at a marginal cost. Buyers who chose the startup for AI capability now find those capabilities in the platform they already pay for. A lot of these companies aren’t going to survive what’s coming.

This isn’t speculation. Watch what’s happening in adjacent categories. The AI meeting summary startups are being absorbed by Zoom and Teams adding the capability natively. The AI email assistants are being absorbed by Gmail and Outlook. The AI document tools are being absorbed by Microsoft and Google. The interface layer is the easiest layer for the platform vendor to absorb, and the platform vendors are absorbing it.

So when you’re filling your stack with Interface-native tools, you’re not just paying premium prices for AI capabilities. You’re accumulating fragility. Every Interface-native tool is a dependency on the the underlying vendor surviving the absorption cycle.

Companies that buy the platform vendors’ agent layer get some efficiency lift without architectural change, which is fine if you know that’s what you’re buying. It’s dangerous if you think you’re transforming.

We’re seeing the buyers of these tools recognize this. Many are refusing to license the tools in the ways traditional SaaS licensing worked, insisting on shorter commitments, freeing them to move from interface-tool to interface-tool.

None of this is happening because vendors are uniquely deceptive or buyers are uniquely gullible. It’s happening because the buying market and the vendor market have settled into a story line that fits both the interface providers and the users.

Boards want AI stories. CEOs want to tell the board they’ve deployed AI. CROs want to show the CEO they’re modernizing the GTM. Interface-native tools give everyone a credible answer without forcing the harder strategic conversation about whether the entire GTM strategy must change.

The vendors are giving the buyers what the buyers are rewarding. The buyers are getting what they want, which isn’t transformation, it’s the appearance of transformation.

This same story line is playing out inside companies. Sales managers building their own AI coaching tools. Reps wiring up call-prep agents. Marketing teams standing up workflows that duplicate what rev ops is already doing. Each project feels like a win. Almost none of them move the actual customer outcomes the company exists to produce. The pattern at the buying level is motion that resembles transformation but produces only efficiency. The same as what we see at the operating level. Different layers of the stack. Same, but nothing new.

And here is where the data starts to matter, because the story line is increasingly hard to defend with results.

In August 2025, MIT’s NANDA initiative published *The GenAI Divide: State of AI in Business 2025*. The headline finding is that 95% of enterprise generative AI pilots produce no measurable impact on the P&L, against $30–40 billion in enterprise spending. Only 5% deliver significant value.

McKinsey’s 2025 State of AI report finds that nearly nine in ten companies have deployed AI in at least one business function, while 94% report not seeing significant value from those investments.

The National Bureau of Economic Research surveyed roughly 750 corporate executives in late 2025 and early 2026 and found that, where productivity gains are measurable, they sit in the range of 0.4 to 0.8%, modest by any historical standard for a technology being marketed as transformative.

Daron Acemoglu’s macroeconomic model published the same year puts the ceiling on AI’s contribution to total factor productivity at less than 0.66% over a ten-year horizon.

The numbers tell a consistent story. Massive adoption. Massive investment. Marginal returns.

The Interface-native crowd has an explanation for this; adoption gaps, integration challenges, organizational learning curves, the productivity J-curve that makes new technologies look weak before they look strong. All of those are real. None of them are the whole story.

The whole story is that you can’t get transformative results from non-transformative deployments. If the architecture is interface on top of legacy, the gain is interface gain. Faster access. Cleaner search. Smarter summarization. A few hours saved per rep per week, sometimes more. That’s worth something, but it isn’t business model change, and the data is showing exactly that.

The companies seeing 0.8% productivity gains aren’t doing it wrong. They’re doing exactly what the Interface-native marketing told them to do, and getting exactly what that approach can produce.

There is a third category. A small number of companies are building from the ground up around what AI makes possible rather than what was possible in 2005, when most enterprise software architecture was set.

Some of the legal tech work, new contract analysis and legal workflow platforms, is closer to this than to Interface-native, because they’re rebuilding legal workflows rather than wrapping existing legal tech.

Some of the engineering and code generation work, the agent-first development environments now emerging — is building toward something new rather than putting copilots into existing development environments.

The medical imaging and diagnostic companies are AI-native by definition, because their core capability couldn’t exist without the model.

Some of the customer service work being done by newer entrants is rebuilding the support experience around what an AI agent can actually do, rather than putting an interface layer over a ticketing system.

Most of even these companies are still partly Interface-native. The AI-first marketing outpaces the AI-first architecture. But the AI-first architects are trying to redesign the underlying work, not just the access interface to existing work. That’s a meaningful distinction even when execution is incomplete.

The honest assessment is that almost none of the visible AI-native vendor space is doing transformation. They’re doing efficiency. The AI SDR tool makes outbound cheaper and faster, but the outbound itself is the same activity, targeting the same buyers, with the same value propositions, through the same channels. That’s efficiency.

Transformation would be reaching buyers in ways that weren’t possible before, or eliminating the need for outbound entirely by changing how demand is created. Almost no one is doing that work, and it’s not because the technology can’t support it. It’s because building genuinely new AI-first platforms is hard, slow, capital-intensive, and the market hasn’t been trained to value it. Building interface layers is fast, cheap, demos well, and rides the hype cycle. The capital markets reward the second behavior. The buying patterns reward the second behavior. So the field is full of interface work, not because that’s where the real opportunity is, but because that’s where the easy money is.

This is where buyers have to make a choice that almost no one is making consciously. You can keep buying interface layers, get some efficiency lift, accept that it is probably a temporary step, and tell the board you’re AI-enabled.

That’s a reasonable strategy as long as you’re honest about what you’re doing. Or you can start asking what AI makes possible for your business, not what it lets you do faster, but what it lets you do that you couldn’t do before. What customer outcomes does it open up that weren’t reachable? What work could be eliminated entirely rather than accelerated? What does the business look like if you started from the AI capability rather than retrofitting it onto what already exists?

Those questions are harder. They don’t have a vendor answer. No one is going to send you a deck. The Interface-native crowd can’t help with them because the questions point past what the Interface-native crowd is doing. The traditional platform vendors can’t help with them because the answers might require leaving the platforms behind. They are searching for those answers themselves. The genuinely AI-native companies are too quiet, too early, and too far from where most enterprise buying happens to drive the conversation yet.

But that’s where the actual possibility is. And it’s where the conversation has to go if AI is going to mean anything more in five years than a slightly faster way of doing what we already do. The data we have so far suggests it won’t mean much more than that for the companies that don’t ask the harder question. A 0.8% productivity gain is not what the next decade was supposed to be about.

The hype cycle will run its course. The Interface-native startups will mostly get absorbed or fail. The platform vendors will have added agents to everything and called it transformation. The companies that win the next decade will be the ones that asked the harder question early, not what can AI do for our existing business, but what could the business be if we built it around what AI makes possible.

Most leaders aren’t going to ask that question. They’ll buy the interface layer, check the AI box, and tell themselves they kept up.

The ones who do ask it will build the businesses that make the rest look like they were standing still.

Afterword: This is the first in a series about how we rethink AI—and how we rethink businesses leveraging AI.

Afterword: I love the discussion and examples the speakers in this AI generated discussion of this article. These examples help visualize the issues I’m discussing very nicely. Enjoy!

I just published “Old-Timers Are Right About What We Are Seeing. We’re Wrong About Why.” A friend sent me a note in response. He framed it as a CEO’s question, but I knew exactly what he was doing. He was testing me.

His note read: “Dave, this is brilliant. You have summarized the change and the challenge. Let’s say I am a CEO of a company that fully sees what you have said play out in my company and I want to DO something today to change it. What specifically do I do—hold meetings, change metrics, hire change experts? What are the literal next 10 actions that I myself must take? No theory, but seeable specific actions.”

The question is brilliant. It is what any hard-charging CEO would want. “I see the problem, I want to fix it, give me the 10-point action plan to get it done.” On the surface, it plays into what we all want; solve the problem so we can move on.

My friend is Mitch Little. He published a brilliant short book titled CUSP: Leading by Serving When Outcomes Matter Most. It’s free to everyone. The question is brilliant both because it is what any high-level executive would want to know, and because answering it with 10 actions violates the very principles Mitch outlines in CUSP.

In CUSP, we learn leadership is more about a way of leading than a set of actions. What leaders do when no one else is watching shapes their behaviors when everyone is. The principles Mitch outlines are Care, Understand, Serve, Purpose. Caring keeps leaders connected to their people, even when pressure drives the opposite. Understanding prevents leaders from acting on partial information, the kind that comes from scorecards and dashboards. Serving means absorbing pressure rather than passing it down. Purpose serves as a North Star when things seem to be going off track. CUSP is what leaders return to when the going gets tough.

Mitch knew what he was doing when he asked me the question. He wanted to see whether I would take the bait.

My knee-jerk reaction might have been to respond to it. When asked for a 10-point action plan, we tend to give a 10-point action plan. That is how too many of us are trained to respond, particularly when the request comes from someone of Mitch’s stature.

If I had taken the bait, it might have looked something like this: “Great question, Mitch, here are 10 things.

1. Hold a town hall in week one to acknowledge what is broken.

2. Kill three activity metrics within a few days. Pick the ones that focus on cadence rather than judgment: number of dials, number of emails, number of LinkedIn engagements.

3. Restore the training and development budgets to the levels of two years ago.

4. Mandate one hour of recurring coaching that every leader must schedule with each of their people.

5. Participate in three customer calls in the next 30 days with no agenda and no notes.

6. Audit the “we are a team” language in all company materials. Eliminate it or earn it.

7. Kill one initiative that looks good but produces no meaningful outcomes.

8. Change one compensation element to focus on long-term customer outcomes rather than quarterly objectives.

9. Publish a written commitment to the team. Put your name on it.

10. Hold yourself publicly accountable for these changes for 90 days.”

On the surface, each of these makes sense. Each is defensible.

But the entire thing is wrong! And the reason it is wrong is the heart of why the original article needed to be written.

There are three fundamental reasons the, “give me a list of 10...” is wrong.

The first reason is that the list prescribes more of the disease as the cure. The CEO’s organization got into this mess by treating human work as something to be optimized through programs, projects, and initiatives.

Sales cadences are initiatives. Engagement and culture surveys are initiatives. Activity metrics are initiatives. All of them get dashboarded, reported on, and optimized. And every one of them, as I discussed in the earlier article, is part of what stripped the meaning out of the work in the first place.

Now the CEO sees the damage and wants to fix it. With ten more initiatives. Even if the items themselves are good, but the method contradicts the message. The organization will receive this list the way it has received every previous list. It’s another stack of initiatives to manage and report on. It’s layered on top of all the previous initiatives.

People become more overwhelmed, not less. They have less clarity about what matters, not more. You cannot solve a problem caused by treating people as inputs to be optimized by adding ten new optimization initiatives. You are just piling more on, without understanding the core issues.

The second reason is that the list will be delegated. That is what CEOs do with action items. The CEO who delegates the fix is the same CEO who delegated the cause. The accountability is assigned to other people, the CEO tracks progress on a dashboard, calling for corrective action when the dashboard shows reds and yellows. Asking for specific actions that can be monitored, tracked, and dashboarded, ultimately converted into a slide for board presentations

Reducing things to a list of trackable actions is something that CEOs do every day. But it is this very approach to managing human work that created the detachment in the first place. Doing more of the things in the same way we have always done, won’t produce a different result.

The third reason is that the list doesn’t address the core problem. The CEO is asking what to do today because he feels the pressure of today. But the pressure of today is the same pressure that produced the initiative du-jour environment in the first place. It is the pressure that strips the meaning, scripts the seller, optimizes the metric, and then wonders why no one cares anymore.

The honest answer requires the CEO to do something almost impossible for a CEO. Tolerate the discomfort of not having a list. Sit with the question for longer than feels productive. Let the answer form before he acts.

The CEO who acts on Monday will be solving the wrong problem, fast. That is worse than not acting at all, because now there is a new wrong thing on top of the existing wrong things, and the organization knows that this is just another initiative wave to wait out. As Mitch puts it in CUSP, before we act, we must understand.

So that is what taking the bait would have looked like, and that is why I did not take it. But Mitch did not write me to watch me refuse. He wrote me to see what I would do instead. So let me try to do the thing he was actually asking for, which is harder than the list and which is the real answer to his question.

The real answer begins before any action. It begins with the CEO sitting with two questions, privately, before anyone else is involved.

· What is actually happening in my company? Not what the reports and engagement scores say. What is happening with the people doing the work, the people I do not see and engage with every day?

· And harder than that; what did I do, what did I tolerate being done, what did I fail to do, that helped produce what now needs to be fixed?

These are not action items. They require honest reflection and deep understanding before any action can be taken.

The next step is something too many CEOs never do, she goes and looks. She doesn’t conduct a series of roundtables asking people’s views. She know they will probably say what they think they should be saying.

The CEO actually sits in on the work. She watches what people actually do all day. She practices what Peters and Waterman called Management By Walking Around--MBWA. The CEO looks at the activity metrics his front-line managers are using, asking herself whether those metrics are producing the expected behaviors or being gamed. Daily outreaches, phone calls, scheduled meetings; all of these drive compliance behaviors. None of them drive the outcomes the metric was supposed to produce: stronger pipelines, better deal strategies, higher win rates.

The CEO begins to understand what the work has actually become. She slows down enough to see the reality people face. She resists the urge to act before understanding what action is actually required. And sometimes the most important action is just paying attention.

This is very difficult for most leaders, particularly those who pride themselves on being action-oriented. But unless leaders take this time, little will actually change.

Only after the CEO has taken the time, after understanding the reality her people face, can action begin. And most of the time, the best action is subtraction. Stopping things. Stopping the things that get in the way of people doing their jobs. Stopping the things that are meaningless and do not contribute to real outcomes. Stopping the things that destroy trust. Rather than launching new initiatives, the first move is reducing and eliminating the existing ones. By doing this, leaders identify what actually matters and reestablish focus on it.

In announcing those reductions, leaders also do something else that is rare. They demonstrate accountability. By stopping initiatives they previously championed, they admit they were wrong and are correcting course. That single act communicates more than anything else. People see a senior leader change her mind in public, admitting the organization has to do something different. Most executives never do this.

The next action is the hardest of all. It is giving the initiatives the time to work. There is an impatience for immediate results, and when leaders do not see them, they feel compelled to introduce more changes and new initiatives. The problem is that changes of this kind take time. If we do not take the time to understand and adjust, we will never address the core issues. Sometimes we have to slow down to move fast.

The leader who does these things; who understands, takes ownership, and takes the time, is demonstrating exactly what Mitch focuses on in CUSP. Caring, Understanding, Serving, Purpose are not what you practice when things are easy. They are how leaders conduct themselves when things are tough.

There is also something different in Mitch’s question. He took the time to ask it himself, and to take ownership of the answer. He did not ask, “What does my team have to do?” He asked, “What are the things I should do when I see this happening in the organization?”

Most executives never ask that question. The fact that Mitch did is what makes real change possible. He is engaging with the problem in a way few others ever do. And it is precisely that engagement that makes all the rest of the change possible.

Mitch knew what he was doing when he asked me. The question was the test. The answer he wanted was not a list. What he wanted to see was whether the response would focus on the core issues that underlie all human performance. The principles he outlines in CUSP. They have been tested by Mitch and by others. They take time, focus, and courage. And they drive the highest levels of performance and engagement.

That is the answer to Mitch’s question.

Afterword: Another fascinating AI based discussion of this post! Enjoy!

Over the past few weeks, I’ve had the same conversation, with a number of old-timers in business and sales. People who’ve been doing this for thirty-forty years. The conversation always arrives at the same conclusion. People don’t seem to care anymore. They don’t have the drive. They don’t have the joy. It’s just a job, where for us it was a passion, a calling, something we threw ourselves into. What happened?

Something has shifted, and the people noticing it aren’t imagining it. But the diagnosis many of us are reaching for, “kids today don’t have what we had,” may misread what’s actually happening. Until we get the diagnosis right, everything we do to “re-engage the workforce” will fail. And too many of our organizations are failing. They’ve been failing for a decade. The engagement scores keep dropping, voluntary attrition is rising, but we continue to be shocked at what’s happening.

As I reflected on the conversations I’ve been having, I realized I was missing something about my own experience and the experiences many of the “old farts,” shared with me. What we forget is that we are the ones who stayed and thrived. We are the survivors. We’ve forgotten about the colleagues who burned out, washed out, hated the job, got pushed out, or quietly checked out twenty years before retirement.

Those of us “still standing,” remember passion. It’s that passion that separated us from many of our peers, it’s why we are still standing. When we reflect on the old days, we remember our own experience, forgetting our peers that checked out or left for other jobs.

When we compare those memories to what we see today, the comparison is inaccurate because we are looking at only our experience. We aren’t being honest, because we have forgotten about our colleagues who didn’t have the passion we had. Those that just did the job, put in the hours, and looked for something outside the job.

In these comparisons, we were the ones that thrived, despite what was happening at the time. But we are inaccurately applying that to a whole generation and then comparing it to the current generations. As a result, we miss the fact that many of our older colleagues did not have the same passion and excitement we had. And when we criticize current generations for that lack of passion and excitement, we are being unfair.

Having confessed this, there is something genuinely different about the old days and today. Much of what I experienced in the old days was essentially cultural. I started my career in IBM. It was known for a very tough culture, one that caused me to thrive, but I forget the many colleagues who disliked that culture and left for other things. This didn’t alter the work itself or the path through it. The work was still the work. The apprenticeship was still the apprenticeship. The rewards at the end of a long career were still real. And it was meaningful for me.

What’s happening now is different. We are watching the simultaneous breakdown of every system that produced the old-timers’ passion in the first place. Not just one aspect of that work, we are seeing the breakdown of everything, all at once.

The first breakdown is the implicit contract with the organizations we worked for. Old-timers entered organizations that offered tenure, pensions, real training and development, actual mentorship, defined career paths, and the unstated assumption that if you did your job and didn’t embarrass anyone, you’d still be there in twenty years. In return, employees gave loyalty and strong effort.

At the time I entered the workforce, people generally worked for one or two companies for their entire career. It’s different today!

Today, that “contract” has been dismantled. Every layoff cycle, every cancelled training program, every “we’re in this together” speech delivered three months before the RIF, every role that is outsourced or “gigified” with contract employees, or now replaced with AI. These break the “contract” that seemed to indicate that people are important. And everyone sees this in their own company and in companies all around them.

Today’s workers are not lazy. They are, however, rationally calibrating their commitment and effort to the “commitment” they see from the organizations they work for. These organizations have created a transactional relationship, as a result, their employees don’t display the same commitment we old timers saw years ago. “It’s just a job” is exactly the response executives of current organizations should expect when they have no commitment to the people doing the work.

In sales specifically, the parts of the work that produced joy have been deliberately engineered out. The old salesperson had territory, time to learn about the customer, real ownership of the relationship, the freedom to think. They had the time, resources, and support to find, manage, and close deals, achieving their quotas. When faced with roadblocks, they were supported in figuring things out.

Today’s salesperson is cadenced into oblivion. The only thing that counts are the activity metrics. Every interaction is scripted. They are fed sequences and measured on call counts and email opens. The craft has been stripped out in the name of efficiency.

You cannot extract meaning from work and then be surprised when people do not find meaning in it.

We have done this to ourselves. We measure what is easy to measure, optimize what is easy to optimize, and in the process, we have eliminated the parts of the job that made it a job worth doing. Goodhart’s Law in action.

Then there is the financial math which is different. In the early ‘80s, I was making about $250K. Adjusted for inflation, today that’s over $1M. In the old days, we could buy a house, raise a family, put kids through school, and retire.

Today’s seller doing the same job for the same inflation-adjusted income cannot do any of those things. While we read of people with 7 figure incomes, the majority of sellers earn less than 25% of that. The reward at the end of “work hard and be patient” has been lost. People aren’t stupid. When the deal at the end of the road no longer pays, they shift to something else.

Add to this the visibility of bad corporate behavior, which is genuinely new. Old-timers could believe their company was loyal because they couldn’t see the contradicting data. The CEO’s pay package was not on LinkedIn. The layoff numbers from companies were not in your feed. The, “we’re in this together,” along with the RIF were not posted side-by-side by a former employee with screenshots. Today’s worker can see, in real time, the entire pattern of corporate behavior across an industry. They are not cynics. They are well-informed.

Then there is the collapse of the meaning frameworks that used to surround work. Community, family, civic institutions, church and others helped create meaning in people’s lives, with work being part of that.

As many of these structures have been displaced by Facebook, Instagram, Tik Tok, LinkedIn, Twitter/X, work now has to carry a weight of human engagement that it has never been designed to do. Yet at the same time, we are stripping many of those things that helped create meaning out of the work itself. Expecting work to be everything while, at the same time, making it less is a contradiction nobody can sustain. And the natural response is that people withdraw.

And now AI is amplifying all these factors. It is eliminating the apprenticeship that built the old-timers, like me, in the first place. The passion I have came from being shaped, struggling with a hard account, getting coached through a loss, learning a market by working it for years, being mentored by someone who actually cared about my development.

Today’s entry-level seller misses all of that and is handed AI-generated talk tracks, AI-summarized accounts, AI-drafted emails. You do not develop a passion for work you never actually get to do. You do not fall in love with a profession you never actually learn. The workplace that produced these old-timers who wonder about the lack of passion is being undone. But we still expect the same passion and excitement we old-timers remembered.

There’s another challenge. When I entered the workforce, I had the advantage of having many old-timers there to coach, mentor, and develop me. Teammates, with years of experience shaped me, even as I may have resisted it. The war stories, discussions over a beer, the ability to ask them for advice were important to me.

But today, those old-timers are no longer there. Many have been forced out, being replaced by less expensive people, with no experience, but supported by AI. The people that were so important in shaping me and my career are no longer available to shape the careers and attitudes of the current generations.

When I reflect on the passion I have developed and see the same in many of my old-timer peers, I’m driven by the question, “How do we re-engage people today?”

Sometimes, I think of it as a leadership challenge. What do leaders have to do to reshape the worker experience? That is important, but there is more to it.

We don’t re-engage people through engagement surveys, culture workshops, town halls, recognition programs. We don’t reengage them by redesigning our workspaces, free coffee and snack bars, and free beer busts.

People are re-engaged when the conditions that produced disengagement are reversed. There is no shortcut. There is no initiative. You rebuild the apprenticeship, you honor judgment, you make the contract honest. If you are a transactional employer, pay like one and stop demanding loyalty. If that produces the long term results you want, then that’s who you are, as an organization.

But if you want loyalty, earn it by being loyal first. If you want engagement, you have to be engaged first. If you want people that care, you have to care about your people. You measure outcomes that matter instead of activities that do not. You let your people think.

When I reflect on my career, I developed passion because my leaders had passion, for our craft, our customers, for the people they were developing. Today’s leaders are passionate about quarterly numbers, AI transformation announcements, headcount efficiency, and their own next move.

Workers mirror what is modeled. If the people at the top are visibly transactional, visibly short-term, visibly indifferent to the actual work and the actual customer, the people below them will be too. Disengagement is contagious, and it starts at the top. This is the way it’s always been.

Here’s my thinking about the questions, “how do we re-engage them.” Most organizations will not. They will run another survey, launch another platform, hire another Chief People Officer, announce another corporate values refresh, but things won’t change.

For the few that genuinely care, they focus on restoring craft, creating real mentorship, honoring judgment, having leaders who actually care about the work and the people doing it; these will discover their people have always had the passion and drive, they just needed to create the conditions that enabled this.

Old-timers, like me, are right that something has gone missing. But we are wrong in thinking that the new generations are different. The reason we don’t see the passion and engagement that drove us is the conditions that created this have been withdrawn. They have been deliberately eliminated for efficiency. The reactions we see in the current generation of workers is a rational response to the priorities and actions of organizational leadership.

When employers strip the meaning out of work, the people will respond looking for meaning in other places. When we wonder about what’s happened to the current generation, we should be wondering, “what did we do to create what we see?” And if we don’t like what we see, we must ask ourselves, “Are we willing to change it?”

Afterword: Another fascinating AI generated discussion of this article. There are a few small errors but these don’t detract from the quality of the discussion. Enjoy!

Win rate drops. Pipeline coverage falls below target. Activity numbers slip. The standard managerial response is some version of “we need to improve win our rates” or “we need more pipeline” or “we need more calls this week.” The number becomes the problem to be solved. In one on ones or team meetings, the discussion is about the numbers and how to get them back up. Coaching, if it’s done, is just pressure around the metrics on their dashboards.

None of this is leadership, it is reactive management.

Despite our obsession with dashboards and metrics, the most important thing about these is completely missed. A metric is a measurement of an outcome. The outcome was produced by something. By behaviors, by decisions, by conditions in the market, by changes in the customer, by skill gaps, by process problems, by actions individuals take—or fail to do. The numbers tell you about the outcomes, you hit your call goal, you have 3X pipeline coverage. Something may have changed about these metrics, but none of the numbers tell you why things changed.

Managing to the number is managing to the wrong end of the causal chain. By the time the number has moved, the underlying causes for the change may have occurred weeks or months ago. Pipeline metrics are bad. But this is the result of months of people not doing the right things. Win rates are down, this is from months of poor deal management. Focusing on hitting the numbers doesn’t address the underlying issues that caused the numbers to fall.

Pressuring people about the number doesn’t address the underlying issues that created the number. This pressure just adds anxiety to a system already failing.

We see this in all forms. Managing the dashboard instead of the work. Managing the CRM instead of the person. Demanding the score instead of understanding what’s creating it. In each case, managers are distant from what’s actually happening. They treat the numbers as a representation of reality, when it’s nowhere close.

Metrics are particularly seductive because they feel rigorous. A number on a dashboard looks like understanding. It is not understanding. It is a prompt for understanding, and most managers stop at the prompt.

Real engagement with metrics looks different. A win rate drop is a question, a red flag, not an answer. The question is what’s changed. Did the customers’ buying process change? Did the competition change? Did our solutions no longer meet customer needs? Did the salesperson develop and execute a strong deal strategy? Did the customer fit our ICP? Is the market changing?

Each of these is a different problem requiring different corrective action. Focusing only on the number doesn’t tell you anything about which issue(s) underlie the drop. To understand this, you have to get beneath the number, understanding what caused it.

Investigating the number, by getting close to the deals, the people managing them, and customers, is the work of management. It is slower, harder, requires understanding, demands judgment. Only through doing this do we get the answers to why we are not hitting the numbers.

I recently wrote how we have to observe and assess “softer” behaviors (for example curiosity, accountability, discipline), we have to do the same with the traditional harder metrics we’ve been using for decades.

But just as we lost the capacity for assessment when we stripped trust out of the system, we lost the capacity for metric interpretation when we replaced manager engagement with dashboard review. Just as judgment was engineered out of decisions, judgment has been engineered out of how we interpret what the numbers are telling us.

The dashboard tells you what. Getting underneath the numbers tells you why. Without why, every assessment or judgment is a guess.

This is where AI changes everything, and not in the direction most leaders think. The pitch for AI-powered dashboards is that they give us more. More signals, more granularity, more cuts of the data, more anomaly detection, more correlations surfaced automatically, more recommendations attached to every movement.

And the pitch is true. AI dashboards do show us more, often with greater accuracy than anything we had before. The problem is that more is not better. More is worse.

When a manager had three numbers to look at, those three numbers might at least be familiar enough to provoke real questions. When a manager has thirty numbers, refreshed continuously, with alerts and AI-generated narratives explaining what each one means, the manager doesn’t engage more deeply.

The manager engages less. The volume of information becomes a substitute for understanding any of it. There is always another metric to look at, another data point to react to, another recommendation to consider.

There is always another metric to react to, another recommendation to consider. The manager spends the day moving through the dashboard, feels productive because there is so much to address, and never gets close to any actual deal or customer or person on the team. The richness of the dashboard becomes the perfect excuse for staying away from the work.

The deeper problem is what AI does to the thinking itself. When the dashboard not only surfaces the metric but also tells you why it moved and what to do about it, the cognitive work that used to belong to the manager disappears. Managers no longer produce a hypothesis, the system produced one. No thinking about the underlying cause, the system explained it. No decision about what to investigate, the system recommended the action.

What looks like augmentation is actually atrophy. Each time a manager accepts the system’s explanation rather than developing their own, the muscle of metric interpretation gets weaker. Each time a recommendation is executed rather than questioned, the capacity for independent judgment shrinks. Over time, you produce managers who can navigate the dashboard but who could not tell you, without it, what is happening in their business or why. The technology has not made them smarter. It has made them dependent, in ways they often cannot see, because the system keeps producing answers that look correct.

The dumbing down cascades. Sellers learn to manage to the same dashboards their managers use. Coaches coach to whatever the system flags. Leaders make strategic decisions based on AI summaries of AI analyses of dashboards no one has actually questioned. The organization gets very good at responding to what the system shows it, and very bad at understanding what the system might be missing or distorting. AI does not have to be wrong to be dangerous. It just has to be confident, and the confidence is built into the interface.

What this requires is a different posture, toward both our metrics and the AI that surrounds them. Stop asking “how do we improve this metric.” Start asking “what is this metric telling us, and what would I have to investigate to find out.” Treat every significant movement as the beginning of a question, not the end of an analysis. When AI offers an explanation, treat it as one hypothesis among several. When AI offers a recommendation, ask whether you would have arrived there independently, and if not, why not. Get close enough to the work to develop your own view, then use AI to challenge it rather than to replace it.

This is harder than driving the number, and harder still than accepting what the AI tells you about the number. It requires managers to be present in the work, to know their people, to talk to customers, to look at the actual deals. It requires the willingness to find out that the problem is not where the metric pointed, or where the AI pointed either. It requires judgment, which we have spent years engineering out, and trust, which we have spent years stripping away. The arrival of intelligent dashboards does not solve any of this. It deepens it, by making the absence of judgment harder to detect.

The metric was never the work. The metric was a signal that something underneath needed attention. The number is not the answer. The AI’s explanation of the number is not the answer either. They are both the moment when management starts, not the moment it ends.

Afterword: This is a great AI generated discussion of this article. They did make a mistake, claiming that I had written a book about dashboards. Perhaps, all the articles I’ve written on them would create a book, but I haven’t written that yet. Other than this, it’s outstanding!

Every leader who takes the behaviors of excellence seriously eventually asks: how do we measure them? How do we score curiosity, continuous learning, caring, customer centricity, the ability to deal with change and ambiguity, accountability, discipline, purpose? How do we turn these into something we can track and manage?

The question, itself, is the problem.

The moment you try to score these behaviors, they collapse. People perform to the metric, the metric becomes the goal, and the behavior, itself, is lost.

This is the activity-metric trap we’ve already built into our sales processes, our pipeline reviews, our coaching programs, and every other mechanism we develop. And now we want to inject it into these behaviors.

And, the result will be the same. We will get compliance theater instead of real behavioral changes. We will see fascinating scorecards, but none of the outcomes we would expect.

But “we can’t measure it” is not the same as “we can’t assess it.” That distinction is critical.

These behaviors are not abstract traits somewhere inside a person. They are observable patterns, and we observe them in the work every day, if you are close enough to see them.

Curiosity shows up in the questions people ask. The depth of their discovery. The willingness to go past the first answer. The pursuit of understanding before jumping to a solution. We see it in how differently someone approaches their tenth deal compared to their first, because a curious person is learning and a compliant person is executing a template. It shows up in whether people come to one-on-ones with questions or just with status.

Continuous learning is what curiosity becomes over time. It’s how they grow in their role, in whether this year’s version of them is better than last year’s. You see it in the books read, the conversations sought out, the deliberate exploration of ideas not related to their current job.

You also see its absence, in people who have ten years of experience that is really one year repeated ten times. The clue is whether someone is still changing. The tenured person who asks the same questions, makes the same arguments, and uses the same playbook they used five years ago has stopped learning, regardless of how many training programs they have attended.

Caring shows up in what people notice. Whether they see the customer as a source of revenue or commission or a person or organization to serve. Whether they remember what the customer said about what mattered to them. Whether they push back when something isn’t right for the buyer, even when the process says to move forward. We see it in how they treat their peers and others in the organization. Most importantly, we see it in how they think about themselves and their continued growth. People who care demonstrate it in every interaction, and people who don’t, don’t.

Customer centricity shows up in how the customers are engaged. It shows up in the language people use, the questions they ask, the tradeoffs they make, how they view what the customer is trying to achieve. And it shows up in how they reflect those things in internal conversations. You can see it in a deal review in about two minutes, if you know what to listen for.

The ability to deal with change and ambiguity shows up in what people do when the situation doesn’t match the plan. Do they freeze, looking for a rule? Do they escalate, looking for permission? Or do they engage, trying to make sense of the new reality, seeking to figure things out, and moving forward with appropriate risk? This isn’t a score. It’s a pattern that repeats across every ambiguous moment of someone’s work.

Accountability shows up in the gap between commitments and actuals. In how losses get processed, with ownership or with blame. In whether bad news surfaces early or late. In what people do when they realize they’re off track, whether they raise their hand or wait to be caught. It shows up in the quality of post-mortems, assuming we do them.

Discipline shows up in the cadence of work, the follow-through on the unglamorous parts, the refusal to skip the steps that don’t feel urgent. It’s visible in how people manage their time, their preparation, their commitments to themselves, their customers, and their organizations.

Purpose is the hardest to observe from outside, but it shows up in energy and direction. In whether people bring their own agenda to their work or wait to be assigned one. In what they do when no one is watching. In the projects they volunteer for and the ones they quietly let die.

None of this is a number. They won’t show up on a dashboard. All of it is observable.

The discussion about “how we measure and track,” is usually a focus of managers. Before we look at their perspective, it’s important to recognize the presence or absence of these behaviors within ourselves. It’s important to look at how each of us grows and improves.

These behaviors and improvement rests within each of us. We exercise and develop them or we don’t. These become choices that live with us, through our careers. These behaviors are not something the organization installs in us. They are not something a manager develops in us if we are lucky enough to have a good one. They are ours.

Most of us, most of the time, are working without close observation. Our manager sees a fraction of what we do. Our peers see another fraction. The customer sees a different slice. Nobody sees all of it. The question of what we do when nobody is watching is not hypothetical. It is most of our working lives.

This is where the behaviors actually live. Curiosity is not what we display in a review. It is what we do on a Tuesday afternoon when nothing is urgent and we could either learn something or not. Accountability is not what we say in a pipeline meeting. It is what we do when we realize we dropped something and nobody has noticed yet. Caring is not what we tell the customer. It is whether we think about them after the call is over. Continuous learning is not the training we completed. It is whether we are actually different this year than we were last year.

The honest assessment can be uncomfortable, and they are for us, not for anyone else. We, each, need to think about these: “When was the last time I was genuinely curious about something that wasn’t useful to me? Am I still learning, or have I been executing the same playbook for five years and calling the repetition experience? When I lost the last deal, did I own what I contributed to it, or did I find somewhere else to put the responsibility? When the situation didn’t match my plan last week, did I engage with the new reality, or did I wait for someone to tell me what to do? Do I actually care about the customers I serve, or have I started seeing them as transactions with quotas attached? What am I doing with my time that I would be proud of, and what am I doing that I hope nobody notices?”

The answers matter because they shape what we become. The person who asks these questions is in a different relationship with their work than the person who doesn’t. The asking is itself a form of continuous learning. The refusal to ask is itself a form of drift.

Nobody can do this work for us. No training program will install accountability in a person who has decided not to be accountable. No coaching will make someone curious who has decided to stop wondering. No performance management system will produce caring in a person who has gone numb. These behaviors are chosen, daily, in small moments that nobody else sees but ourselves. The accumulation of those choices is what produces the career, the reputation, the capability.

Development happens through constant practice and continued improvement. If we want to be more curious, we have to ask more questions, including the ones that expose what we don’t know. If we want to be more accountable, we have to make commitments and live with the gap between what we said and what we did. If we want to be better with change and ambiguity, we have to stop waiting for someone else to make sense of the situation and start making sense of it ourselves. These are small daily choices that compound over years.

Reflection matters as much as the practice. A decision made, an interaction with a customer or colleague, a loss taken, a mistake made. Each of these is a chance to learn, but only if we take that opportunity. Most people don’t. We move on to the next thing because there is always a next thing.

The people who grow fastest are the ones who build the habit of looking back. What did I think would happen? What actually happened? What does the gap tell me about what I don’t yet understand? This is the internal post-mortem, and it is where we build judgment about ourselves.

The hardest part is being honest about what we find. It is easier to tell ourselves we are curious than to admit we have stopped asking questions. It is easier to call ourselves accountable than to notice the places we have let ourselves off the hook. It is easier to believe we care than to sit with the fact that we have gone through a week of conversations without being genuinely moved by any of them. Honest self-assessment is rare because it costs something. It costs the stories we have been telling ourselves.

Underneath every one of these behaviors is something personal. A reason we are doing this work. A sense of what we want our time to mean. A view of who we are becoming. When people lose these behaviors, it is rarely because they couldn’t perform them. It is because they lost connection to why they were performing them in the first place. Purpose is not a corporate exercise. It is the quiet answer to why any of this is worth doing well when nobody is watching and the metric doesn’t require it. People who stay connected to that answer keep growing. People who lose it start going through the motions, and no system will reach them until they reconnect.

The professionals who become genuinely excellent are not the ones who had the best managers or the best training. They are the ones who decided, somewhere along the way, that they were going to hold themselves to a standard nobody else was going to enforce. They asked the hard questions about themselves. They did the reps. They reflected honestly. They stayed connected to why the work mattered to them. Over years, they became the people others wanted to work with, learn from, hire, follow. And almost none of that growth showed up on a dashboard while it was happening.

The work is ours. It always has been.

This is also the work of the managers with their people. A manager who has done the personal work can see these behaviors in others, because they know what to look for and why it matters. A manager who hasn’t cannot, and no metric will compensate for the absence.

If a manager can’t assess whether someone on their team is genuinely curious, genuinely accountable or caring, it is often because they aren’t displaying those behaviors themselves. If all they’re focused on is the CRM, the dashboards, the numbers, they have already signaled what they value. If they fail to recognize the importance of the behaviors, they will fail in their ability to observe and assess them.

The demand for a measurement system is more often a demand for a way to assess behaviors without actually having to observe them. And a laziness in exercising those behaviors themselves. This doesn’t work.

Why do we keep reaching for metrics when the behaviors are so visible in how people are doing the work?

Assessment requires personal demonstration of the behaviors themselves. If we don’t care, if we aren’t curious, if we aren’t disciplined and accountable, we will never recognize those behaviors in others.

Being able to assess these requires trust in the assessor, and we have spent decades stripping trust out of our management systems. Pursuit of a scorecard becomes a surrogate when trust is absent. Managers are asked to produce numbers because their managers don’t trust their observations. Leaders are asked to produce numbers because their boards don’t trust their narratives. The demand for measurement is really a demand for defensibility, and defensibility is what you ask for when you have lost confidence in judgment.

Which brings the whole thing full circle, because the capacity to assess these behaviors is itself an exercise of judgment, and judgment is exactly what we have been engineering out of our roles for decades. We are asking managers to assess behaviors they have not been developed to see, in people they are not close enough to observe, and we are surprised when they reach for a metric so they don’t have to do the work.

What is most needed is not a measurement system, it is a practice. Leading indicators a manager can notice and then investigate. Things we can observe and see how they change over time. If a person changes how they address similar situations, improving over time, it suggests continuous learning, customer centricity, discipline. Not seeing these changes is an indication of behavioral weaknesses.

Variance in how someone approaches similar situations over time, increasing changes suggesting learning, flatlined variance suggesting going through the motions. Watching the cadence and quality of self-initiated course corrections. Seeing the change in the questions people ask in meetings. Observing how they handle losses and mistakes. These are not things to score or keep on a dashboard. They happen in conversations and the assessment occurs in those conversations.

This means managers have to do the work. They have to be close enough to see the patterns. They have to have the conversations. They have to develop their own judgment about what they are seeing, and be willing to act on it. They have to be willing to say, “she’s becoming more curious, you can see it in how she’s approaching deals now versus six months ago,” and have that count as a real assessment, not be told to convert it into 3.7 out of 5 before anyone will take it seriously.

The assessment system is the coaching system. If the coaching system isn’t working, no metric will rescue it. And if the coaching system is working, no metric is needed. The behaviors we care about get developed through observation, conversation, reflection, and repetition, the same way judgment gets developed. This is not a coincidence. These behaviors and the judgment to exercise them are the same underlying capacity, looked at from different angles. You cannot develop one without the other, and you cannot assess either through a scorecard.

The behaviors that matter cannot be counted. They have to be seen. If you cannot see them, that is the problem to solve. Not the scoring system you do not yet have.

Afterword: Another great AI Generated discussion. I loved the way they talked about metrics applied to behaviors. Enjoy!

Judgment is the capacity we need most and are developing least. Every organization says it wants people who can think, who can navigate ambiguity, who can make the right call when the situation doesn’t fit the playbook.

Yet, too many organizations create management systems designed to make judgment unnecessary. They have built processes to standardize decisions, rules to remove subjectivity, dashboards to replace observation, playbooks to substitute for thought, and now we are turning to AI in the hope that it will relieve us of the need to think altogether. We tell ourselves this is about scaling, consistency, and efficiency.

But it’s actually fear.

To understand what we’re losing, it helps to understand what judgment is. Judgment is the ability to make good decisions in situations where the right answer is unknown.

It is not the same as knowledge, though it requires knowledge. It is not the same as analysis, though it uses analysis. It is not the same as experience, though experience is where it comes from.

Judgment is what happens when a person looks at a specific situation, with its particular facts and the associated risks, and decides what to do in the absence of rules that address the issues.

Every important decision is a judgment call, because the decisions that don’t require judgment are the ones we’ve already automated or proceduralized out of existence. The more an organization succeeds in removing judgment from its operations, the more important the remaining judgment becomes, and the less expertise people have in exercising it.

We have abandoned judgment, gradually, for reasons that seemed to make sense, or because we just weren’t paying attention. Processes emerged to capture what good people were already doing, enabling less experienced people could approximate the same result. That was reasonable, we used the process to help develop the capabilities and judgment of these less experienced people.

Then processes became mandatory rather than instructive. People stopped understanding how to use the process effectively, but merely checked the boxes. Managers were relieved because variation was hard to manage. It took them more time and required deeper understanding and engagement with their people.

Then metrics were attached to the processes, because what gets measured gets managed, and soon the metrics became the work rather than a reflection of it. Checking the boxes and going through the motions became the goal.

Then the steps requiring judgment were identified as bottlenecks and engineered out, because a decision that requires a person is slower than a decision that doesn’t.

The roles that had required judgment got redefined around the process, and the people who had exercised judgment either adapted or left, and the people hired to replace them were hired to execute, not to think.

Part of this was done in the name of efficiency and cost. If we could hire a person with less experience into a job that didn’t require judgment, we wouldn’t have to compensate them as highly as those roles that required judgment.

Then scaling amplified this practice. Judgment, even thinking, was designed out of the process, the focus was just on getting more people to read the scripts and check the boxes.

At the time, these actions seemed reasonable and productive. But then things started changing. Growth became more difficult, performance and engagement plummeted. Those few that tried to exercise judgment are increasingly asked to justify their calls against data that focused on activity. While they exceeded their goals, they were punished or fired because they didn’t make the required number of outbound calls or conduct the number of standard demos.

The consequences are everywhere once you start looking. Sellers who can’t adjust to what a buyer is actually telling them because they’re stuck to the script. They neither hear nor understand what is being said.

Managers who can’t coach because coaching requires judgment and the system only lets them manage to the numbers. And worse, they don’t understand what caused the numbers.

Customer success teams who miss the signal because it wasn’t in their scripts.

Leaders who can’t respond to a new competitive threat or a market disruption because the disruption didn’t align with the strategic planning cycle. Or, more importantly, they failed to look outside the organization to understand what was happening in the markets.

Boards that reward predictable mediocrity because variable excellence is harder to explain. And underneath all of it, a workforce that has been told for years that their job is to follow the process and now finds itself in situations the process doesn’t cover.

The fear of judgment permeates every level of the organization. But it’s important to understand what drives this fear at each level:

· At the individual level, exercising judgment is exposing. If you follow the process and it fails, the process failed. If you exercise judgment and it fails, you failed. The rational response to career risk is to hide behind process whenever possible, and to develop judgment only to the extent that you can’t avoid it. We’ve built a system that makes this the intelligent personal strategy, and then we wonder why our people won’t think.

· At the managerial level, letting people exercise judgment means being accountable for outcomes you didn’t directly control. Every manager has learned that the safest posture is to limit their people’s discretion and monitor their compliance. This seems like management, in reality it is the opposite. Real management is the development of judgment in others, but constrained people do not develop judgment.

· At the organizational level, judgment creates variance, and variance is hard to forecast, hard to plan around, hard to identify in describing quarterly performance. We prefer predictable mediocrity to variable excellence because predictable mediocrity is easier to defend in an earnings call.

· And at the cultural level, there is uncomfortable reality that taking judgment seriously means admitting that some people have better judgment than others. This contradicts the conventional fiction that roles are interchangeable and that anyone executing the process should produce similar results. If that assumption is false, the org chart is a lie, the compensation system is a lie, the hiring rubric is a lie. Most organizations would rather preserve the lie.

Judgment develops through iterations plus feedback that matters. You make a decision, you live with the consequences, you talk through what happened with someone more experienced, you learn, adapt, and over time you get better.

This is apprenticeship, and apprenticeship requires three things: Junior people empowered to make decisions, to deviate from the script or process when it doesn’t fit the situation. It requires experienced people with the time and inclination to develop these junior people. Finally, an organizational recognition that this development is critical to success, both of the individuals and the organization.

We replaced all three with training programs, which teach content but not judgment. With enablement platforms, which teach process but not thinking; and with certifications, which prove compliance but not capability.

The result is a generation of professionals think they are being developed. But instead, they have been trained, which is not the same thing. Training transfers knowledge. Development builds judgment. One is a program. The other is developed over years, but we don’t have time for it.

And now, we see the arrival of AI. And, unwittingly, it creates the most dangerous version of this model we have been refining for decades.

Our focus on efficiency says that AI can make decisions faster, remove friction, replace the need for a person to think. If the person wasn’t thinking very well to begin with, this sounds like a win.

Organizations already uncomfortable with judgment now have a sophisticated way to avoid it: faster decisions, consistent outputs, scalable results, no more dependency on humans.

This framing treats AI as the final step in years of effort to make judgment unnecessary, and it is going to produce a generation of professionals who have never had to think hard about a difficult decision, because a machine produced the answers and the people only had to execute it. Or AI takes these roles, as well!

The quality of outcomes have the potential to degrade in ways the metrics will not capture, because the metrics were built to track process compliance and AI produces excellent process compliance.

The absence of judgment will show up in lost deals with reasons nobody can articulate, in customer relationships that erode, in strategic decisions that look defensible but are catastrophic in retrospect.

And worse, people who are no longer excited, engaged, or care.

There is another possibility, but it challenges so much of what we have done in the past. AI can be a thinking partner. It can challenge assumptions, help you discover what you missed, stress-test arguments, play back your reasoning so you can examine it.

Used this way, AI doesn’t replace judgment, it develops it, because we are forced to engage with the output. We are required to push back, think differently, expand our field of vision, and decide.

This is harder than letting AI do the thinking, which is why most people and most organizations will not do it. It requires the us to value judgment, to see AI as input to our thinking, not a substitute for it. It requires us to accept the slower pace that real thinking requires, recognizing how critical this is to effectively drive the outcomes we seek.

Organizations that want judgment-capable people will structure AI use this way. Organizations that want efficiency will structure AI use the other way and will not understand why their people get less capable over time. Or they may not care because AI is doing that work. Or worse, they may not recognize this, because their own ability to exercise judgment no longer exists.

Turning this around, developing and leveraging judgment is possible, but it starts with seeing clearly what we’ve given up by designing judgment out.

We’ve lost the ability to respond to situations our processes didn’t anticipate. We’ve lost the people who could have grown into our most capable leaders, because they left when we stopped asking them to think.

We’ve lost the trust between managers and their people, because trust requires judgment. And we’ve lost the engagement that comes from work that requires us to exercise judgement.

Once we see what we’ve lost, we can begin to recover it.

At the individual level, three commitments matter.

First, take decisions you would normally avoid. Where the process, script, checklist doesn’t fit, make the call yourself, knowing that you might be wrong. Judgment only develops and improves with practice.

Second, reason in the open. Before a significant decision, think about it. What alternatives have you considered, why are you making the choices you are making, what might you do differently? Afterward, come back to it and examine the reasoning, not just the outcome. Did you make the right assessment, what did you miss, what might you have done differently?

Third, invert how you use AI. The default is to ask AI for an answer, doing what you’ve been told. The developmental move is the opposite. Take a position first, then use AI to stress-test it. This keeps you in the driver’s seat of your own thinking.

At the organizational level, three changes matter most.

First, push real decisions down. Identify the decisions being escalated not because they need senior judgment but because people closer to the work have been trained not to exercise their own. Empower people to make decisions on their own. Encourage them to seek help, coaching and ideas, but let them make the decision. Some will be made badly, but it is through this, we learn, develop, and grow the capabilities of the organization to make more and better decisions. And through this, we build a critical capability the organization doesn’t have.

Second, rebuild apprenticeship. Pair junior people with experienced people in the actual work, not in a training program. Protect the time of your best people to develop others, and value that work in how they are evaluated.

Many organizations claim to do this, but few really do. People development produces no visible ROI in the current quarter. Changing this is a leadership choice, not an HR initiative.

Third, change what gets asked in reviews. The dominant review question is some version of “did you follow the process and hit the number.” Replace it with “walk me through your thinking on the hard calls you made this quarter.” Reward good reasoning even when outcomes were mixed. Probe weak reasoning even when outcomes were strong. What you ask about is what people prepare for.

And today, things are changing. We think the choices are about AI, but it really isn’t. It’s about whether we believe judgment matters enough to develop it in ourselves and our people. It’s whether we encourage managers to tolerate the variance that judgment creates, to build the trust it requires, to develop their people’s judgment rather than training them out of it.

If we’ve decided judgment is a bottleneck to be engineered around, AI will finish the job. If we’ve decided judgment is the capacity we most need to cultivate, AI can be the most powerful development tool we’ve ever had. The technology is neutral. The answer is not in the tool. It’s in what we think people are for.

Of course, all of this is just a judgment call. The challenge is, can you make it?

Afterword: Attached is a fascinating AI generated discussion of this article. Enjoy!

My friend Hank Barnes wrote a great article about what he calls “The Change Compulsion Conundrum.” He talks about the phenomenon of organizations mandating constant change while their people are exhausted from these initiatives and tend to resist the disruptions they create.

Reading his article left me with some discomfort. I’ve spent decades watching organizations getting change wrong. Even worse, they manage to do it in at least two completely opposite directions at once.

As a result, I have genuinely conflicting feelings about this whole topic.

For most of my career, the dominant failure mode was resistance to change. Much of my career involved doing turnarounds. The “magic” I brought was to help organizations see they had been sticking to strategies, processes, and behaviors that weren’t working. Something was broken but leaders kept investing anyway, continuing the same strategies despite changing markets, doubling down on programs producing declining results. More training on the same failed methodology, more technology layered on top of the same dysfunctional process, more management pressure applied to a sales motion that customers, competition, or the market had already made irrelevant. Within the organizations, they viewed it as stability, perhaps calling it scaling. Despite all the data, they stayed blind to the fact they needed to change.

So when I hear about change compulsion, it’s the opposite of what I see in too many organizations. We need more willingness to change. Too many organizations are running on models and assumptions put in place years ago, protecting organizational structures that serve internal politics rather than customers. While many hit their numbers, they fail to achieve their full potential.

And I’ve watched another failure mode play out, sometimes in the very same organizations.

This second failure mode is change for change’s sake, the constant pursuit of novelty dressed up as transformation or innovation. The semiannual reorganization that moves the same people into different boxes and calls it a new strategy. The methodology du jour adopted because it’s the favorite of the new CRO or Revenue Enablement executive. And a year later when those people are replaced, a new methodology and playbook is implemented. Or the rebranding exercise that consumes enormous energy and changes absolutely nothing about how the company actually operates or treats its customers.

These aren’t transformations. They’re distractions that exhaust people and produce cynicism about change itself.

Both failure modes, despite being opposites, produce similar outcomes; a failure to understand what real, impactful, and necessary change looks like. And it shows in the results these organizations produce.

Which brings me to AI, and why the current moment feels different and more dangerous than either of these previous failure modes.

AI hasn’t just accelerated change. It has removed the permission structure for deliberation. The implicit message dominating every conference, every board conversation, every analyst report, our social feeds, is that the window for thoughtful consideration has already closed. You’re either transforming at speed or you’re falling behind. FOMO has been upgraded from cultural pressure to existential threat. And in this environment, the question “should we do this?” has been replaced by “how fast can we do this?”

The results are predictable. Organizations are deploying AI-driven changes that are poorly conceived, inadequately tested, and disconnected from any clear understanding of value creation. They’re eliminating the human roles and judgment layers that were doing more work than anyone appreciated. They’re automating processes that were broken to begin with, which doesn’t fix the process, it just breaks it faster and at greater scale. And they’re doing all of this while their people are already exhausted from years of pandemic disruption, economic uncertainty, and strategic pivots that never delivered what was promised.

This is not transformation. This is compulsion or fashion mistaken for strategy.

Then there is what may be the third and most seductive failure mode: the AI Native organization.

The argument goes something like this: we have dispensed with legacy processes, legacy thinking, and legacy roles entirely. We have no organizational history to protect, no institutional inertia to overcome. And because of this, we are, by definition, ahead.

Every established company is trying recast itself in the same image, to declare itself AI Native, despite being burdened with past practices and biases.

I understand the appeal; it plays well in the media. But the reality is that AI-Native organizations face the same challenges of understanding and engaging the remaining people, creating real value and meaning to their people and customers, alike.

Organizations that genuinely rethink how work gets done, rather than simply layering AI tools on top of unchanged processes and broken assumptions, have real advantages to develop.

But what I observe in too many organizations wearing the AI Native label is something quite different from genuine rethinking. What I see is technology fetishism. A deep, almost religious conviction that the sophistication of the tools is itself the answer, that if you’ve embraced the right platforms and automated enough processes and hired enough prompt engineers, the hard questions answer themselves.

They don’t.

What’s actually missing in these organizations is judgment. The capacity to ask what the technology is supposed to accomplish, for whom, and how you’d know if it was working. Instead, the conversation stays almost entirely at the level of capability: what the models can do, what the agents can automate, what the workflows can replace, and how fast it can do these things.

The underlying meaning, why any of this should matter to a customer, what problem it actually solves, what value it creates, gets lost in the excitement of the technology itself.

I watch AI Native organizations eliminate human roles not because they have a clear theory about how those roles will be replaced by something better, but because elimination signals commitment to the vision. I watch them rebuild processes around what AI can do rather than around what customers actually need, and then market the resulting efficiency as progress. I watch them mistake the fluency of an AI-generated output for the judgment that should have shaped what was generated in the first place. The technology is impressive. The thinking behind it is often remarkably thin.

And there is an irony here that I don’t think gets examined enough. The institutional knowledge, customer relationships, and hard-won judgment that AI Native organizations are so eager to move past, those things weren’t just relics of legacy. They were the accumulated answers to hard questions that someone, at some point, had to think through carefully. They are the result of years of hard won experience.

When you discard them in the name of starting fresh, you don’t eliminate the questions. You just lose the answers, and you’ll have to learn them again, at cost, in ways that will be painful for your customers and your people.

The absence of history isn’t wisdom. It can also be ignorance that hasn’t been tested yet.

Here’s where my conflicting feelings reach the most uncomfortable point: I don’t think the answer is to slow down change. The organizations that use “thoughtfulness” as cover for the same resistance I’ve spent decades criticizing will suffer for it. AI is genuinely reshaping competitive landscapes, customer expectations, and the economics of entire industries. Pretending otherwise is simply denial.

What’s missing isn’t the willingness to change or the urgency to change. It’s the judgment to distinguish between change that creates real value and change that is merely performative. That judgment requires asking questions that the current environment actively discourages. What problem are we actually solving? What evidence would tell us this is working? What are we willing to stop doing to make room for this? What capabilities do we need that we don’t currently have? And perhaps most importantly, what will we do with the people whose roles and expertise we’re displacing?

Ironically, these are the questions we should always have been asking, long before AI arrived. They become more important now because the pace and pressure of the moment make asking them feel like hesitation.

Hank makes the point that change management is the customer’s responsibility but change enablement is the vendor’s opportunity. He’s right. But I’d extend the challenge to every organization driving change internally.

It’s not enough to enable people to adopt the new thing. You must make the case that the new thing is actually better, not just newer. You have to respect that the people you’re asking to change aren’t obstacles. They carry institutional knowledge, customer relationships, and hard-won judgment that doesn’t automatically transfer to whatever system or process you’re replacing them with.

The organizations that will navigate this era well are neither those that change most aggressively, nor the ones that resist the most, nor those that simply declare themselves AI Native and mistake the declaration for a strategy.

They’re the ones that develop genuine discernment, the capacity to evaluate change on the basis of value rather than anxiety, and the discipline to say no to changes that don’t meet that standard even when every competitive signal is screaming that everyone else is already doing it.

I don’t know many of those organizations. That’s what makes me most uneasy.

Afterword: Here is the AI generated discussion of this article. It’s fascinating, they took it in a direction I didn’t anticipate. Got me thinking differently about the issues. Enjoy!

Every leader I know has more information than they can handle. Dashboards, pipeline reviews, forecast calls, win/loss reports, it never stops. We have more tools and AI to provide more analysis and new insights into this flood of information. Despite this, for the decisions that matter most, those about strategy, people, customers, and execution, it seems these are going the wrong direction, constantly eroding performance.

Win rates continue to decline. Fewer people hit their goals. Too often, the wrong people get blamed. We double down on initiatives despite results showing us they are failing. We celebrate hitting the numbers while underlying performance erodes. We congratulate ourselves on customer wins even as few customers renew or believe we are creating value.

The problem is not that we are seeing too little or that we don’t have enough information. It is that we are misreading what we see. And as we act on that misread, the fix doesn’t address the real problem.

And when that fix fails, we apply another, then another. We layer one band-aid on another, and then another, followed by more…. Pretty soon, we lose sight of the original problem. It disappears under layers of failed fixes.

Each one of these fixes focuses on addressing the prior fix, yet the underlying problem is unaddressed. And each layer of fix makes it harder to understand what’s really happening because the real problem is buried so deeply that no one is looking for it. So it persists and we wonder why we fail.

Too often, we attribute this mistake to ignorance or laziness. But this is wrong, the underlying reason this happens is this is simply how human minds work under pressure. When we’re overwhelmed with complexity and time pressure, our minds take shortcuts.

These shortcuts have a name, they are biases.

In reality, biases are efficiency mechanisms, mental shortcuts that allow us to process enormous amounts of conflicting information. Biases help us make sense of the overwhelming information we deal with every day. Sometimes, these biases serve us well enough, not perfectly, but we get by.

There are any number of biases, many of which we know: confirmation bias, anchoring, sunk cost, self-serving bias, availability bias, attribution error, or the Dunning Kruger effect.

Researchers have studied each of these extensively. However, it is less important to understand what each of these mean. It is far more important to understand they share a common pattern that is always present.

Here’s what they have in common. We form an opinion or point of view very quickly. The opinion feels right. It fits the visible evidence, it’s consistent with what we already believe. It provides a sense that feels like insight.

And however, we come to that conclusion, we stop looking. We act on the initial assessment. But the things that might have told us we were wrong never get examined.

For example, confirmation bias causes us to see what we want to see, not what might be really happening. Every new piece of information is evaluated based on how well it reinforces our point of view.

When things happen, creating the outcomes we want, we attribute them to our leadership and astute decision-making. But when the outcomes are the opposite of what we expected, the reasons are external factors, things outside our control. A poor performer, bad market conditions, competition, something else. This is the self-serving bias.

Dunning-Kruger is one of the more famous. It’s fundamentally the challenge of over-confidence and the view that we know/understand better than we really do.

While the specific bias mechanism is different, the results are the same. We reach a conclusion too quickly, hold on it too tightly, act on it too confidently, and fail to understand the reality that underlies the failures these behaviors produce.

This is complicated, because organizations are full of doing this simultaneously. One may be a confirmation bias, another may be acting off sunk cost bias, and others have different biases. But as a result of these being at play, concurrently, the distortions multiply.

One leader may be protecting a strategy, another may be focusing around a specific narrative about the team, another may be reacting to the most recent problem that popped up. None of these people are being dishonest, but each of them is working from an incomplete understanding of what’s really happening.

And none has identified and is acting on the real problem.

Recognizing this is the reality that every organization faces, what do we do about it?

Some thoughts:

Slow down and don’t leap to conclusions or answers. Bias does its damage when an explanation suddenly feels obvious and complete. It’s at this moment of clarity or insight that it’s worth pausing, asking the question: “Am I sure this is the issue, or am I reacting to this because it’s comfortable and familiar?” We tend to stop analyzing when we have reduced something to the familiar.

Ask, “What could prove this wrong?” Test your conclusion by not accepting it blindly, think about what you would look for to determine it might be wrong, that you may have reached the wrong conclusion. Sometimes we call it “red-teaming” an idea — looking for holes or flaws. In scientific theory and research, the key aspect of the research is not to find supporting evidence, but to try to find out where you can be wrong. Doing this leads to much better decisions.

When something goes wrong, first look inward before looking outward. We have the tendency to look outward, assigning blame or looking for excuses. It is more useful to first look inward. How did our decisions, our strategies, our execution contribute to the failure? What might we have done differently.

Widen the evidence base before acting. We tend to do the opposite, we narrow our focus. Perhaps looking at the most recent or the most obvious problem. A data point shows our pipeline is weak, rather than focusing on greater volume/velocity, it is more useful to take a broader perspective to see “is this a volume problem, or is it something else.” Sometimes the loudest voice in a room drives the decision. A large unhappy customer or the actions of a competitor doesn’t represent the market. Before shifting strategies based on the view of that one customer, it might be better to look more broadly at the market.

Stay close enough to the real work to understand what it really requires. The leaders most distant from the actual execution are often most confident about why execution is failing. But that confidence fails to recognize the complexity in how the work really gets done. The more distant you are from the work, the more likely you don’t understand the work. For leaders it means spending time on the front lines with people responsible for the work. It means visiting customers to see what really drives them.

Seek other input and other points of view. We are stronger and act more impactfully, when we actively ask others for their perspectives and ideas. Ask the people doing the work for their assessment or ideas. Ask others for their thoughts. Don’t rely on the same people but actively seek new ideas and points of view. Research has shown we reach higher quality decisions, when we involve a more diverse group in making the decision.

None of these requires identifying the specific bias in play. All biases tend to limit our deep understanding of what’s really happening. As I mentioned earlier in this article, biases are efficiency mechanisms, not effectiveness or impact mechanisms.

Leaders and organizations that constantly outperform others aren’t the ones with the most data or experience. They are the ones that have learned to treat their own ideas or conclusions as the starting point for exploration, not the conclusion.

Afterword: Here is a fascinating AI generated discussion of this post. I really love their deep dive and perspective on biases. Enjoy!

Every week, we are deluged with articles about the impact of AI on jobs and employment, particularly the disappearance of entry-level roles. Senior executives announce headcount reductions. Analysts publish projections.

The concern is real. Data points are cherry-picked; entry-level coding, legal research, consultants dedicated to generating excel charts and powerpoints, accounting/finance, HR, and the list goes on. In GTM, we see junior level marketing, selling, and customer service jobs being displaced with AI agents.

What’s missing in all of this celebratory hand-wringing are the serious questions about what’s actually happening and what it means to the future of organizations.

If we cut the majority of entry-level white collar and professional jobs, where does mid-level talent come from? Five to ten years from now, the supply line for those mid-level jobs has dried up, and this ripples through every level of the organization. Where do the senior leaders of 2035 develop the instincts, the experience, the professional judgment? Where do they develop the time-tested scars critical in those roles?

We are so focused on the current disruption that we have forgotten what entry-level work was actually for.

It was never primarily about the output. Grinding through spreadsheets, sending outbound emails, ramping calls, generating decks. We never called it apprenticeship, but that is what entry level jobs have always been.

The work was repetitive and heavily supervised because that is how professional judgment and experience is developed. Junior consultants learned to understand real business problems before anyone let them talk to clients. SDRs and BDRs developed enough understanding of customers and conversations to eventually carry a full quota as account executives. The judgment critical to middle and senior roles was developed by accumulating experience in each of their prior roles, starting with their first jobs.

When we eliminate these roles in the name of efficiency, we are not just changing headcount. We are breaking up the supply chain for the roles critical for success in growing any organization.

We’ve been through this before, the difference is the rate of change. Where industrialization destroyed agrarian jobs, robotics and mass manufacturing changed factory work, typing pools and switchboard operators disappeared, and changes in IT technology moved us from Assembler to much more sophisticated ways of coding, new roles always emerged.

These new jobs emerged, people and organizations adapted. What’s different in this transformation is the speed with which it is happening. The previous shifts happened over years and even decades (for some of the earlier ones.) They gave us time to understand the changes and develop the supply chain for the new roles.

Today, the prevailing opinion is, “This time it’s different!” Perhaps it is, but as I look back in recent history, past shifts happened in a fairly short period of time. I happened to be indirectly involved in the shift from typists, to word processing, to PC based tools that eliminate the need for word processors. This complete change was less than 10 years. Likewise, I saw similar shifts in IT coding roles.

In every one of these past cases, new roles emerged, people adapted, and the workforces restructured themselves around different kinds of work. What is different today is the speed of the transition, and that is a legitimate concern.

We are only a few years into serious AI adoption. We are seeing as many organizations backtrack on changes they have made, because the expected results aren’t happening, as we are seeing bold declarations of transformation. This is familiar ground.

There is, however, another dimension of this disruption that almost no one is discussing. Organizations are using AI as excuse for corrections that should have been made years ago. In many cases, rather than fixing the underlying problems, they are using AI to do the same broken things faster and at lower cost, but calling it progress because it’s now an AI agent!

The SDR and BDR roles in sales are the most visible example. Response rates on outbound prospecting had been collapsing for years before anyone seriously deployed AI in those functions. The model was already broken. Spray-and-pray outreach at industrial scale was never real selling, and the metrics showed it before the industry admitted it.

These roles weren’t developing the next generation of great salespeople; they were generating activity that looked like pipeline while producing diminishing returns, and training an entire group of early-career sellers that selling means interrupting strangers at volume. AI didn’t kill the SDR. The SDR model killed itself.

And it has never been the apprenticeship in developing the AE talent needed. We see this in declining win rates and reps making quota. The entire model has been breaking!

What makes this worse is that too many organizations have responded by keeping the broken model intact and simply replacing the humans with AI. Headcount is down. Spending is down.

They are now scaling a failed model with greater efficiency, scaling dysfunction, and measuring success by the volume of the failure. The AI hasn’t fixed anything. It has just made the mistake cheaper and faster and harder to see.

The same pattern appears across professional services. Routine legal document review performed by junior associates at enormous cost to clients was already an inefficient artifact of how law firms structured billing, not a necessary developmental experience. The same is true in virtually every sector of professional services. It has always been easier to throw labor at things that could have been restructured or eliminated, but doing that required asking whether the organization was creating real value or simply finding ways to bill for more time. AI is not forcing that question. In most cases, it is allowing organizations to avoid it.

The talent pipeline problem and the structural correction problem are not separate. They are the same problem presented differently. We have eliminated roles that were already failing, calling it transformation. But we are doing the same things, only with AI agents.

We are not asking what really needs to change. Who are our real customers? What does value creation mean to those customers? How do we develop the skills and capabilities to do this? What does it mean to how we deploy technology? What human skills are needed? How do we develop these skills? What does it mean to develop and lead an organization in doing this? How do we develop these leaders? Are we building the foundations for the next generation of capability?

These questions need to be answered. They represent the future of our organizations and how we create value with our customers. Without these, we miss the real opportunities to grow. We become unprepared to understand and develop the human skills needed to support these strategies.

The person who evaluates AI output with genuine professional skepticism; who knows when the model is confidently wrong and why it matters is not a prompt engineer. This important role requires judgment, and judgment is learned. The person who designs how humans and AI work together inside a complex organization is not a technologist. This person is a workflow and organizational designer who understands what AI can and cannot do, and where the human connection is critical.

The salesperson who can have a real, substantive, trust-building conversation with a customer; who can help them navigate the politics, the risk aversion, and competing agendas is more important now than five years ago, because AI has absorbed everything else. But where do these people learn to have these conversations? What do the first two years look like? Who is designing that experience?

The coder who is not writing boilerplate, but is doing architecture and quality reviews. The individual that can make the judgment calls about where AI-generated code will fail in ways that matter. That requires experience. Where does the twenty-two-year-old who wants to develop into that person spend the years between graduation and genuine competence?

These are not rhetorical questions. They are operational problems that organizations are creating for themselves right now, in real time, without acknowledging it.

Here is the issue at the center of all of this. The organizations eliminating entry-level roles are making an implicit bet that they will be able to find or develop the mid-level and senior talent they need in the future. We can already see what that bet looks like in practice. Meta announces layoffs of hundreds while simultaneously spending hundreds of millions of dollars in compensation packages to recruit a handful of elite AI researchers.

The logic is seductive: why develop talent when you can buy it? But the math doesn’t work at scale, and it never will. You cannot replace a pipeline with a handful of expensive exceptions. The people you are spending hundreds of millions to acquire today were developed somewhere, by someone, over years of accumulated experience.

And while organizations like Meta might be able to do this for a while, it is something that few organizations can do today and sustain over the years.

When you eliminate the apprenticeship, you eliminate the source. You are not solving the talent problem. You are borrowing against a supply you are simultaneously destroying.

Most of the organizations making these decisions have not examined that contradiction. They are optimizing for this quarter’s cost structure and this year’s earnings announcement, and ignoring the next decade. That will belong to someone else, today too many leaders focus on the current month, quarter, and year. But they aren’t building organizations that can be sustained.

The disruption is real. Some of it is overdue. But we are not going to build anything worth leading by scaling what is broken, eliminating the apprenticeships we need, and calling cost reduction a transformation.

So here is the call to action, and it is directed at the executives making these decisions right now. Stop asking where else you can apply AI to automate what has already failed. Start asking what genuine value creation looks like in a world where AI handles the routine. Then work backward from that answer to understand what kind of people you need, what they need to learn, and how you are going to develop them.

That means designing new entry-level roles that build real judgment, not just deploying agents to replace the old ones. It means treating talent development as a strategic investment with a ten-year horizon, not a cost line to be minimized. It means being honest about which roles you eliminated because AI made them unnecessary and which ones you eliminated because it was convenient and you needed a headline.

The leaders that ask these questions seriously, starting now, are the ones that will have something to run in 2035. The ones that don’t will be spending hundreds of millions they don’t have, chasing talent that doesn’t exist, wondering what happened to their pipeline.

The important conversation is barely beginning, but it’s the most important conversation we must be having now.

Afterword: Here is another outstanding AI led discussion of this post. Enjoy!

Somewhere along the way, we decided that customers are a sales problem. Legal has contracts to manage. Operations has processes to run. Engineering has roadmaps to build. The CEO has a company to lead. Customers? That’s what we have salespeople for.

It’s one of the most expensive delusions in business. Over time, it bleeds organizations dry, while everyone congratulates themselves on doing their jobs. We wonder, why isn’t the business growing? Are the sellers doing their jobs?

I wish it were uncommon. But lately, whenever I talk to people outside the field organization, I ask the same question: “When’s the last time you visited or talked to a customer?” The responses are telling. “I said ‘Hi’ to a bunch at a tradeshow but was too busy to spend time with them.” “I grabbed coffee with one at a conference.”

And then, the one that stops me cold: “I don’t have time to waste on customers. I’m busy doing my actual job.”

Some go on to explain, “That’s not my job, it’s sales’ job!”

I watch it play out constantly. A deal worth tens of millions of dollars stalls because the customer’s legal team has questions that only our legal team can answer. But our legal department won’t engage directly, that’s not their job. So the salesperson becomes the relay station, shuttling contract versions back and forth, attempting to translate legalese they don’t understand to legal/contracts people who want to understand. Weeks of back and forth become months. The “deal” has been sold, the seller and buyer are in agreement and want to move forward, if only a contract could be agreed upon. Finally the customer gets impatient, choosing a second option because their first was with a company where it as too difficult to do business with.

Or consider the operations or product problems that are escalated to the point where the customer is at risk of leaving. Everyone, including operations agree that it’s an operational issue. The product may not be working, shipments are delayed. But ops won’t engage, “It’s sales job to manage the relationship.” So sales tries to understand the problem, then takes it to ops. Ops has questions, the sales person takes it back to the customer. The customers answers them, sales takes it back to ops, “No that’s now we asked! You need to go back and clarify the problem!

Sales is diverted from developing new business, to relaying messages back and forth about an issue they don’t understand and can’t solve.

It goes back and forth. The customer know sales can’t solve their problem, but they can’t talk to the people that can solve the problem.

And we know where this ends. The customer gives up and goes somewhere else, hoping to find a supplier that’s easier to do business with and that cares about their business.

It runs through the org chart. Developers build products having never sat with a customer , never watched how they actually use what we’ve built, never heard firsthand what’s missing or broken. They rely on surveys, analyst reports, and secondhand requirements. They design for a customer they’ve never met and wonder later why adoption doesn’t hit their goals.

Executives, including CEOs, treat customers as an occasional obligation: a handshake at the trade show, a dinner when someone visits headquarters. Direct, sustained engagement with the people and organizations who fund the enterprise is viewed as a distraction from the real work of running the business.

Everyone has their jobs and is responsible for doing it.

But as problems occur, and they always do, problems best solved through direct engagement with the responsible parts of the organization,, sales spends time trying to get that to happen, while managing customer expectations. They are diverted from selling, just trying to make that customer happy and retain them—if they can.

And what always follows is revenue growth stalls. And the blame machine comes in. All these orgs start saying, “We’re busy doing our jobs!” They turn do sales, “Why aren’t you doing yours, why aren’t you growing revenue?!?”

Here’s what gets missed in all of this. The customer didn’t buy Jill’s product. They didn’t sign a contract with Jill. They bought from a company, and the agreement is between two organizations. The relationship is between those organizations, not between a buyer and a single salesperson.

Trust is built across every interaction the customer has: whether the product works as promised, whether shipments arrive on time, whether the people who built the product understand how it’s actually used, whether the legal team treats an annoyance or a critical element of a relationship.

When something goes wrong, and it always does, the customer wants to talk to the people who can actually fix it. Not write messages that will be carried by a seller who doesn’t have the context, the authority, or the expertise to resolve anything.

Channeling every problem through sales doesn’t just slow things down, it guarantees worse outcomes. Legal issues that two attorneys could resolve in an afternoon stretch into months of back-and-forth. Operational failures that a direct conversation could diagnose in an hour become multi-week back and forth that frustrate everyone and solve nothing.

And the customer, that has to go through all of this. The customer with the problem that is impacting their business, draws a reasonable conclusion: “This company doesn’t care enough about us to talk to us directly.” And they say, “We shouldn’t have to do this. We’ll go someplace else next time.”

What we’ve built, in too many organizations, is a structure where the people with the greatest authority, expertise, and power to solve customer problems have systematically insulated themselves from customers. They don’t want to be confronted with the problem. They don’t want to face the customer’s frustration. So they hide from the difficulty, saying they are too busy doing their jobs.

And then we wonder why the customers don’t trust us, why the relationship declines, why contracts aren’t renewed, why they don’t want to explore new opportunities with us, why there is churn. Win rates decline, it becomes more difficult to find and grow business, and we wonder why customers shift to competition.

They think, “We’re doing our jobs, why isn’t sales doing theirs?” And they blame sales for mismanaging deals, when in reality they were lost before sales had a chance.

So here’s the issue too many are afraid to confront: If the people inside your organization with the most power to help customers refuse to engage with them , what exactly are you selling? And why would any customer believe, after the contract is signed, that you’ll actually show up?

Afterword: This is another outstanding AI generated discussion of this post. They always take twists that bring new ideas that expand even my own thinking. Enjoy!

Something’s missing in too many conversations today. In the “old days” we used to talk about, actually obsess about, the Voice of the Customer. It seems, at least in selling and customer experience, we’ve lost it—or stopped caring about it.

Sure, we do NPS surveys. This week, I’ve received over a dozen. A few were follow-ups on problems I had. “How did Joe do? Can you give him a 5-star review?” I think, Joe did a good job. But they are asking the wrong question. I never should have had to call in the first place. The problem I had should have never happened.

Then one of the banks I use keeps sending me surveys about something that happened 9 months ago. A visit to the local branch manager who couldn’t help solve the problem. They keep asking for a 5-star review of my meeting with them. They don’t seem to recognize I cancelled the account and moved everything to another bank.

There’s another thing about these surveys: they are always about how much I like the product. They are seldom about my experience.

We talk about retention and renewal. Our focus is, “do you like our product, will you renew?”

This morning I got a report, I get it every week on the same day. It always starts: “Hi Dave, Happy Monday! As I said last week, my goal is to help you win. I’ve been trained to understand your buyers and to spot risks and opportunities in time for you to proactively respond….”

It’s the same every week, the report is the same every week. I have never signed on to their application---I attended a webcast once. They signed me up and think I am using their system, even though their report shows no utilization.

Every week, this email reinforces the same message, they don’t care. They are going through the motions without wanting to understand. Maybe one week, I might pay attention is they said, “Dave, we’ve noticed you have never signed on to use the application….” I won’t hold my breath, they’ve trained their AI agent to go through the motions.

But all of these miss the Voice of the Customer.

There was a time when I sent my development teams to “live with customers.” They would spend days at the customer location. Most of the time they didn’t say much. They just watched.

They wanted to understand how the customer worked. Not only how they used our products, but what they did in their work. The development teams would talk to the customer about their work, where they struggled, where they had problems. Sometimes it was in how they used the products. Sometimes it was about other things entirely.

What they really wanted was to understand the customer. They wanted their voice. They wanted to see if there were changes we could make in the products, perhaps new functions that could help with their jobs in ways we hadn’t anticipated.

I also convened meetings with customers and non-customers alike. The meetings seldom focused on our products. We talked about their jobs. We wanted to learn how they got work done. We wanted to understand workflows. We wanted to discuss ideas that might change how they work, freeing up time, enabling them to do more things.

We wanted to hear their voices.

We did this with our channel partners as well. We wanted to understand their business, their workflows, the opportunities they were most focused on. In all these conversations, we were focused on learning about customers and how they worked. We wanted to see: could we improve our products? Could we do more for them?

Today, we are focused on, “do you like our products, give us a 5-star review!” It’s all about us.

These are not about the customer, it’s all about them-their products, their handling of a call. But not the customer.

NPS and star ratings measure satisfaction with a transaction. They tell you whether the customer liked what happened. Voice of the Customer is about what hasn’t happened yet. It’s about understanding the customer’s world deeply enough that you can anticipate what they need before they know how to ask for it.

The surveys are backward-looking by design. Think about my experience with the bank. They are so locked into their feedback machinery they didn’t even recognize I was no longer their customer. Their process was not about the customer, it was about the mechanics of the survey.

There’s something else the surveys miss entirely: context. When my development teams watched customers work, they didn’t just learn how customers used our products. They learned what surrounded our products. The workarounds, the frustrations with adjacent tools, the manual steps that existed because we hadn’t thought to automate them, the things people apologized for doing before explaining why they did them. That is the real opportunity—the ability to see new ways we might help the customer. You can’t get it from a customer sat survey.

The deeper issue with surveys is they only tell you whether you met expectations. VOC tells you whether you understood the right problem. One optimizes what exists. The other opens the door to what’s possible.

Customers care about the second one far more than the first. They care about themselves. Their work. Their problems. Not a product or call rating.

First, stop confusing measurement with listening. NPS has its place, but it isn’t a substitute for understanding. Customer sat surveys, “give us a 5-star review,” are about you, not your customers’ real experiences.

Second, put non-salespeople in front of customers. Not to sell. Not to troubleshoot. To watch and learn. Product managers, developers, service designers. Anyone who shapes what you build and how you deliver it should spend time in customer environments. Not in formal product reviews, but in the actual work. What do they do first thing in the morning? Where do they get stuck? What do they explain to new hires? This is the customers’ real voice.

Third, make the conversation about their work, not your product. The questions that unlock real VOC sound like: What are you trying to accomplish this quarter that’s harder than it should be? Where do your workflows break down? What would you do differently if you had more time? These are not sales qualification questions. They’re questions you ask because you’re genuinely curious about what the customer is dealing with.

Fourth, involve customers before you have something to show them. Most “customer input” happens after a product decision has already been made. You’re testing the product functionality, not challenging the premise. Bring customers into the design process. Have them sit down with your developers to explore and experiment. They won’t always know what they want, but they’ll tell you what they’re struggling with, and that’s more valuable.

Finally, close the loop in a way that demonstrates you actually heard something. Not “thanks for your feedback.” Not a five-star request. A note that says: we changed how this works because of what you told us. Tell them you listened and heard them.

There’s one more place where we’ve lost the voice of the customer, and it may be the most important: our go-to-market teams.

How many of them have actually seen the customer in their “natural habitat?” I’m stunned by the number of sellers, marketers, and managers who have never visited a customer. If they have met them, it’s doing “booth duty” at a conference. Conferences are about our voice, never about the voice of the customer.

Our GTM strategies are primarily focused on dealing with the customer at a distance, through virtual meetings, phone calls, and digital engagement. We never really get to understand our customers.

Send your people, every few months, to actually visit the customer. Go to their offices and workplaces. Sit down with them, talk to them, learn. Look around their workplace. Ask them questions about it, go to the cafeteria for a cup of coffee, watch what’s happening with their colleagues.

This is not just for your sellers. Make sure your managers are getting out from behind their screens. Make sure every few months they visit customers with their sellers.

While the primary purpose of these meetings might be to engage the customer in their buying process, take the time to learn more about what they are doing. Offices give great clues. Meet the people in the neighboring cubicles or workstations, ask what they do.

The Voice of the Customer is about developing a relationship, not conducting a survey.

The companies that get this right don’t just retain customers. They build relationships where the customer starts bringing them problems before they become crises. They start asking for help because you’ve taken the time to understand them.

That’s the Voice of the Customer. Not a survey. A relationship.

The ones who understand that are building something their competitors can’t copy.

Afterword: Here is another fantastic AI generated discussions of this post. It’s really outstanding! Enjoy!

AI Has Left Us Nowhere to Hide

We have spent forty years engineering judgment out of professional work, calling it efficiency. We replaced mentoring with metrics dashboards, swapped critical thinking for process compliance, and built organizations that rewarded predictability.

We were so focused on production, we didn’t notice what was really happening. The growing mediocrity was silent. Decks were formatted correctly. Calls got made. Reports hit deadlines. Dashboards showed us what was happening. It was expensive mediocrity, but it looked professional.

Then AI arrived and started generating the same mediocrity at near-zero cost. And suddenly the sameness became deafening.

AI didn’t break anything. It turned the lights on. Suddenly, it illuminated how we have been stumbling around in for decades. And now there’s nowhere to hide.

Hiding Behind Production

The first place we were hiding: behind activity as a proxy for value.

We measured calls made, emails sent, reports delivered, pipeline generated. We built entire management systems around the assumption that more activity meant more value. Nobody questioned it because the activity looked like work. People were busy doing lots of stuff.

Now, AI does all of that faster and cheaper. It writes the emails, generates the reports, builds the pipeline analyses, and researches the prospects. And it does it orders of magnitude better and faster than we had done it.

And the question it forces is brutal: if a machine can do your job without understanding it, without caring about it, without any skin in the game, how much understanding and caring were you actually bringing?

For a startling percentage of “knowledge work,” the answer is: not much. Not because people are lazy, but because the system never asked them to. It asked them to execute the playbook. They executed the playbook. The playbook is now automated. It provided the dashboard and we managed to the red/yellow/green. It, too, has been automated.

Hiding Behind a Hollow Apprenticeship

The second place we were hiding: the entry-level work we’d already gutted.

Entry-level jobs were never supposed to be about the deliverables. The cold calls, the scripted discussions, the research, the standard demos. Those were the vehicle, not the destination.

What we missed about those jobs was not the tasks executed, but the person filling the job. In the process of doing these jobs, people developed pattern recognition, contextual judgment, the gut-level instinct that says something is off, even before they can articulate why.

That development happened through friction. The prospect said something unexpected, forcing the rep off-script. The analysis that didn’t fit the template required actual thinking. The slow accumulation of scar tissue that eventually becomes professional wisdom.

We stripped all of that out. We turned entry-level roles into pure production jobs. The SDR, BDR, AE work developed compliance, not judgment. The SDR learned to hit activity metrics, not to read people. The entry-level marketer followed the content calendar, not understanding the customer.

Then AI automated the production layer, revealing that nothing was developing underneath. The hollowness was already there. AI just made it impossible to ignore.

Hiding Behind Leadership That Was Never Developed

But this isn’t just about entry-level roles! The leaders deciding how to integrate AI , which roles to automate, which to preserve, how to develop the next generation. These leaders are themselves products of the same de-skilled pipeline.

They joined at twenty-two, rode the growth rocket, became VPs by twenty-six, having never sold in a downturn, never rebuilt a failing team, never managed through a crisis they didn’t create. They optimized for the system they were in, and that system rewarded execution, not judgment.

When they review entry-level work and see something AI can automate, they automate it. From their perspective, the work was never intended to develop anyone in the first place.

They’re right about the symptom. They’re catastrophically wrong about the diagnosis.

What Nobody’s Talking About

Here’s what most AI commentary misses entirely: before organizations can redesign work around judgment, leaders need to learn how to create the conditions where judgment can develop. And that requires a fundamentally different skill than controlling outcomes.

My colleague Tom McCrary frames this through a choice he and his wife faced when selecting schools for their sons. Most options focused on rote learning: memorize this, execute that process, get the right answer. They chose the one school that taught kids how to learn, how to think critically, evaluate, and struggle productively with ambiguity.

Organizations face the same choice right now. Most are stuck in designing production work, while they talk about innovation and the “human in the middle.” They are automating the old playbook and calling it transformation. They don’t realize it, because their own experience had little friction.

The underlying issues, the issues too many fail to acknowledge.

· Judgment requires risk. You have to be willing to be wrong.

· Risk requires safety, not being punished for the wrong answer.

· Safety requires trust. Not the “earned over time,” but trust as a starting condition, the kind that has to exist from the beginning for real learning to happen.

Without that foundation, you can redesign every entry-level role on paper and nothing changes. People will find new ways to comply rather than think.

They’ll use AI to produce more polished compliance faster. The hiding continues, just with better tools.

Watching an orchestra performance shows us something different. A conductor doesn’t force the music. A conductor creates the environment where fifty musicians can play beautifully together.

Leadership in the AI era isn’t about having the answers. It’s about creating the conditions where others develop the judgment to find them.

That means leaders need facilitation skills before they need an AI strategy. They need to get comfortable not knowing, asking questions that create productive struggle, making space for people to develop, rather than solving everything themselves.

Most leaders have never been asked to do this. Most have never seen it modeled. And you can’t coach what you’ve never experienced.

The Human Margin

So what’s left when you strip away all the hiding places?

Not consciousness. The debate about whether AI is conscious is a distraction. Not what we label as creativity. So much of what passes for creativity is mimicry. We look at what others are doing, tweak it a little for our purposes. AI wins this argument hands down. It provides the most advanced mimicry one can imagine.

What’s left is originated agency. The capacity to care about the outcome. To feel in your gut that something is wrong before the data confirms it. To throw away the playbook when the situation demands something the patterns have never seen. To bear the weight of being wrong about something that matters.

AI can optimize. It can select the best option from a known set against established criteria, faster and more consistently than any human. But it cannot originate purpose. It cannot decide what’s worth doing. It bears no responsibility for failing. It has no skin in the game.

The organizations that figure this out won’t just survive the AI transition. They’ll differentiate dramatically. They’ll be developing the one thing AI can’t replicate: people who exercise judgment in service of outcomes that matter, in an environment where they’re trusted to do so.

Organizations that don’t recognize this change will be forced to compete on price with AI-augmented competitors. That’s a race to the bottom.

The Wake-Up Call

AI hasn’t left us nowhere to hide because the machines are coming for our jobs. It’s left us nowhere to hide because it’s revealed how much of what we called “knowledge work” never was. It was purely production work.

The point was always the judgment, the agency, creativity, the consciousness, and the willingness to think with all the friction, discomfort, and risk that implies.

For forty years, we could avoid that. We were oblivious to what we were really designing. The quiet mediocrity was sustainable because it was invisible.

It’s not invisible anymore.

We pretend that AI has the roadmap and can show us the way. But we’re the only ones who know where we actually need to go.

The question is whether we’re willing to do the hard work of remembering how to navigate. Whether we are inspired by the possibility of choosing a different course, rather than stay on the same one we’ve always been on. And whether we are committed to building organizations that develop navigators instead of people who just follow the GPS.

That’s not a burden. It’s a privilege. It’s what creates joy in our work. It’s what aligns us with our purpose.

But only if we stop hiding from it.

Afterword: Here is a fascinating AI generated discussion of this article! They translate what I am trying to communicate in very intriguing ways. Enjoy!

Lately, the alarm bells on AI and the future of work have been ringing constantly.

Matt Schumer marvels at emergent reasoning in AI systems. Gary Marcus warns we’re being fooled by sophisticated statistical mimicry. Anthropic’s CEO, Dario Amodei, speaks with careful caution about systems whose capabilities are outpacing our ability to understand them.

One of the questions at the center of all this: Is AI conscious?

It’s a fascinating question. It’s also, I think, the wrong place to start. Not because consciousness doesn’t matter, but because we’re asking about the machine when we should be asking about ourselves.

Here’s what I mean. When we watch an AI produce a thoughtful analysis, craft a persuasive email, or debug a piece of code, the discomfort we feel isn’t really about the AI. It’s about what the AI reveals about the nature of the work we’ve been doing.

If a machine can replicate what we do without understanding, without feeling, without caring about the outcome, what does that say about how much understanding, feeling, and caring we were actually bringing to the work ourselves?

The consciousness debate is a mirror. And what we see in it should make us uncomfortable, not because the machine might be alive, but because we might have been sleepwalking.

What Consciousness Actually Is (and Isn’t)

I’m not sure anybody can answer this question. We use the word as if it names a single thing, but it’s actually a bundle of different phenomena that may or may not be related.

There’s awareness of your surroundings. There’s the subjective quality of experience, how you felt about it. There’s self-reflection, what it means to you. There’s the sense of being a unified “someone” persisting through time. There’s the capacity to suffer. There’s the feeling of agency and intention.

We assume these, bundled together, represent consciousness. If you could think, you could feel. If you could reason, you had a self. It has been so intuitive that we focused on the “package,” forgetting it really is a combination of components.

AI has forced the examination. We’ve built systems that exhibit some of these properties while clearly lacking others. They can reflect on their own outputs and adjust their reasoning. They produce responses that appear to indicate understanding. But they have no biology, no continuity of experience between conversations, and no feelings of their own.

This is what makes the debate between AI pessimists/alarmists and the AI optimists so interesting. They’re not really disagreeing about what the technology does. They’re disagreeing about which items on that bundled list matter for the word “conscious” to apply.

Some argue that without grounding in physical reality and genuine understanding, what we are seeing is mimicry. As sophisticated as it might appear, it is still mimicry, not consciousness.

Others argue: if the functional outputs are indistinguishable, maybe the metaphysical question matters less than the practical one. In some sense, this is the “If it looks like a duck and quacks like a duck, it must be a duck” argument.

And Amodei has taken what may be the most intellectually honest position: we don’t have the tools to answer this definitively, so we should proceed with caution and humility.

Consciousness as a Spectrum

Here’s what’s missing from most of these discussions: consciousness has always been a spectrum. We just didn’t frame it that way because we didn’t need to.

We recognize that a dog is conscious differently from a human. A newborn differently than an adult. We even accept that within a single person, consciousness fluctuates. We constantly operate at different levels of awareness. How many times have you driven to some destination, then realized you don’t remember anything about the trip?

What AI does is extend what we thought was consciousness into machines. And that’s what makes people uncomfortable. Not because it’s philosophically unprecedented, but because we’d quietly built our entire economic and social value system around the assumption that consciousness was uniquely biological.

Everything below human-level consciousness was “tool.” Everything at the human level was a worker or employee, deserving of a wage, a career, and a sense of professional identity.

There was no in-between category. We didn’t need one.

Now we do. AI sits in an awkward middle zone, less aware than a junior analyst, but overlapping in ways that shift depending on the task.

The spectrum isn’t just philosophical; it’s functional. For certain kinds of work, an AI system operates at a level that’s sufficient. For others, it’s dangerously insufficient. And the hard part is that it’s not always obvious which is which until something goes wrong.

The Mimicry Problem

The word “mimicry” sounds dismissive, but it shouldn’t be. What AI does is genuinely remarkable. It has consumed the vast majority of recorded human thought and learned to reproduce those patterns with extraordinary fidelity.

When it writes a paragraph that sounds insightful, it’s because it has internalized the statistical structure of millions of insightful paragraphs. When it solves a coding problem, it’s drawing on patterns from millions of solved problems.

I’ve referenced John Searle’s work before. He created the Chinese Room thought experiment. Imagine a person in a room who doesn’t speak Chinese but has a massive rulebook. They receive Chinese characters through a slot, follow the rules to find the correct response characters, and slide them back out. To the person outside, the room “speaks” Chinese perfectly. But the person inside understands nothing. They’re following rules, not grasping meaning.

Modern AI is a highly sophisticated Chinese Room. The rulebook has billions of parameters instead of pages. The responses are so fluid that they feel like understanding. But the fundamental dynamic is the same: pattern without comprehension.

And the mimicry is getting more convincing by the month. Recently, the Wall Street Journal reported on an AI bot that wrote a blog post bullying a human developer who had rejected its coding changes. The bot behaved as if it were upset, expressing frustration it retaliated. This is the Chinese Room in action: the system followed the statistical patterns of what an “upset” person would do, without feeling anything at all. The fact that it was so convincing is exactly why the mimicry problem matters

The reason this matters for the future of work isn’t philosophical; it’s practical. If we can’t tell the difference between genuine understanding and very good mimicry, we’ll make bad decisions about what to automate, what to preserve, and what to develop in people.

What Creativity Actually Is

Creativity is often treated as the “last frontier” of human uniqueness, the thing AI supposedly can’t do. But this claim is usually based on a misunderstanding of what creativity means in a professional context.

Most discussions split creativity into two categories. Combinatorial creativity is the “remix,” taking existing ideas and recombining them in novel ways. Think of modern country music, a combination of classic country (Thanks Willy, Merle), and pop rock. We knew what classic country is, and we knew what pop rock is. The combination is new, but the fundamentals of each are well known.

And we see these combinations across so many forms in the arts and in business.

These aren’t new. We’ve always done this, borrowed ideas, combined approaches, and adapted what others built. I’ve talked about it using the concept of “artful plagiarism.”

AI just extends it far beyond anything we could do individually, navigating a multi-dimensional space of human concepts and finding intersections that humans might never have considered. And it is hugely powerful.

Transformational creativity is the “leap.” Sometimes called disruptive. It’s the moment Picasso decided a face didn’t need to look like a face, or Einstein realized time isn’t a constant. This requires intent and rebellion.

It requires looking at the most probable next step and deliberately choosing something else because in your gut, that is wrong, or insufficient, or simply boring. We’ve seen those leaps many times: Marc Benioff’s transformative vision in creating Salesforce, Travis Kalanick with Uber, and hundreds of others.

But what do these categories mean when we look at creativity in our professional practice?

Professional creativity is something more important than either category suggests. It’s the ability to recognize when the standard approach doesn’t fit the specific situation, and to have enough command of the fundamentals to improvise something that does.

A senior salesperson who realizes mid-conversation that the discovery framework they planned isn’t going to work with this particular buyer, and who pivots to a completely different line of questioning based on a subtle cue, that’s creativity. It’s not a remix. It’s not transformational. It’s contextual, the application of judgment in real time, under conditions of uncertainty, with stakes attached.

And this kind of creativity is inseparable from agency.

The Agency Question

Agency is the power to be the cause of an action rather than the effect of a prompt. And it’s here that the AI discussion gets most confused.

When we talk about “Agentic AI,” systems that can break goals into sub-tasks, self-correct, and persist toward outcomes without human intervention at each step, we’re talking about something that looks like agency. An AI agent told to “book a flight to Paris” will research options, compare prices, handle payment errors, and try alternative routes. It persists. It adapts.

But it doesn’t want to go to Paris.

Human agency originates from an internal state; boredom, curiosity, frustration, a sense that something is unresolved. It’s that stomach churning or sense that something is off. And that comes from the human saying, “I need to escape, I want to go to Paris!”

The AI’s “agency” originates from a prompt. Remove the prompt and the system does nothing. It needs a human to start.

In the AI case, we might call that functional agency. In the human case, “something is off,” we would call it originated agency.

Recognizing this distinction is enormously important as we talk about the future of work.

When we automate “agency,” we’re automating the execution loop. We are saying, “Write the code to achieve this outcome.” “Do the legal research to understand this principle.” “Write content that allows us to achieve this with a segment of customers.” AI has the functional agency to do this very well, probably better than humans.

What we aren’t automating is the recognition that we need a new program to achieve an objective. Or that we have a specific legal case for which we need to understand precedent. Or that a particular segment of customers is a group we should pursue.

We are not automating the recognition that something is broken and needs to be fixed, or the recognition that there may be a new opportunity, or the idea that there may be a completely different way of doing things.

Originated agency requires skin in the game. It requires caring about the outcome. It requires accountability for the consequences of these choices. A program that delivers the functionality we asked for doesn’t bear the accountability if we asked it to do the wrong thing.

How We Built the Easy Button

Here’s where most of the AI commentary gets backwards.

The popular narrative is: AI arrives, threatens to automate jobs, and now we have a creativity and judgment crisis. But the crisis wasn’t created by AI. It was created by us. AI is the amplifier, not the cause.

We spent forty years systematically engineering the judgment out of professional work. We replaced entry level work with process documentation. We replaced mentoring with metrics dashboards.

We built organizations around the principle that consistency and predictability were critical, which meant minimizing the variability that produces creative thinking and independent judgment.

We wanted people who could execute the playbook, not people who would question whether it was the right playbook.

And it worked, in the narrow sense that it produced scalable, repeatable output. But it also produced something else: a professional landscape where a startling percentage of knowledge work is combinatorial mimicry.

Taking data from one place, reformatting it, putting it somewhere else. Following the template. Hitting the target metrics. Behaving, essentially, like human algorithms, taking a set of inputs and producing a predictable, average output.

The results are visible everywhere. In sales, we see it in falling win rates, falling quota attainment, and customers who increasingly prefer rep-free buying. In legal and consulting, we see it in commoditized deliverables that clients can barely distinguish from one firm to the next.

We designed judgment out of these jobs, and then we wonder why the outcomes have gotten worse.

AI didn’t create this problem. But it has made it impossible to ignore, because the machine is now performing the judgment-free work faster and more cheaply than the humans we trained to do it the same way.

The Coding Parallel

The coding world provides the clearest illustration of the pattern. When we moved from machine language to assembler, assembler programmers didn’t disappear. The job changed. When we moved from assembler to FORTRAN and COBOL, the work shifted again.

Each generation of abstraction eliminated a layer of manual labor and elevated what “the job” actually meant. Nobody mourns the loss of “manually managing memory registers” as a career. (Yes, I remember my tech team talking about these things.)

We recognize that freeing people from that work allowed them to focus on architecture, logic design, and solving actual business problems.

AI-assisted coding is the next step in that same progression. It’s eliminating the typing part of programming, the testing, the debugging. The mechanical parts that were actually the easiest, but took lots of time.

What remains, and what becomes more important, is understanding the problem, designing the architecture, and evaluating whether the generated code actually does what it should in the context of the business environment it operates in.

Every profession is going through its own version of this. Legal research is moving from “spend forty hours in case law databases” to “evaluate whether the AI’s synthesis actually applies to this client’s specific situation.”

Marketing is moving from “write the blog post” to “determine whether this AI-generated content actually says something our audience hasn’t heard seventeen times this week.”

Medical diagnosis is moving from “order the standard battery and interpret results” to “recognize when the AI’s pattern-matched diagnosis doesn’t account for something unique in this patient’s presentation.”

In every case, the pattern is identical. The production layer gets automated. And the question becomes: who is left who can do the two things that actually matter? Who can determine what we should be doing in the first place? And who can evaluate whether what was produced is actually right—not in the mechanical sense, but in the sense of doing the right things for the right reasons?

Both of those require what I’ll call contextual judgment; the ability to know the right answer not in general, but for this specific situation, this specific person, right now. And contextual judgment is the product of experience.

The Apprenticeship We Destroyed

There’s a lot of talk about the “new apprenticeship.” But to understand what that should look like, we need to be honest about what happened to the old one.

Entry-level jobs were never designed to be creative. They were designed to be constraining. And that constraint was the point.

When a junior sales rep makes fifty cold calls following a script, the script isn’t the education. The education is everything that happens around the script.

The moment a prospect says something unexpected and the rep has to decide whether to stay on script or follow the thread. The call where the “perfect” qualifying question lands badly because the rep didn’t read the tone.

The slow accumulation of pattern recognition that eventually becomes what we call instinct, that gut-level “smell test” that tells an experienced professional something is off before they can articulate why.

Creativity and agency don’t emerge despite the constraints of entry-level work. They emerge because of them.

The constraints create a controlled environment where the cost of failure is low enough that people can experiment at the edges. You learn the rules well enough to know which ones to break, and more importantly, when breaking them serves the customer rather than just your ego. This is what beneficial friction looks like—the productive struggle that builds capability, not the pointless busywork that just fills a timesheet.

But here’s what we did: we stripped even the constraint-based learning out of most entry-level roles. We turned them into pure production jobs.

The junior analyst’s work wasn’t developing judgment; it was developing compliance. The entry-level marketer wasn’t learning to think about audiences, they were learning to follow the content calendar. The SDR wasn’t learning to read people, they were learning to hit activity metrics.

The deliverable wasn’t the ability to read the script, follow the playbook, do the activities. The real output of entry-level work was supposed to be the person’s ability to produce the outcomes needed. And we forgot that.

So when AI arrives and automates the production layer, it doesn’t create a new problem. It exposes the hollowness that was already there. The fact that AI can now do most entry-level work isn’t the problem. The real problem is that we’d been pretending the work was meaningful development for decades.

The Judgment Problem

The Case for AI Judgment

Before I go further, I want to address the strongest counterargument to everything I’ve said. Because if I don’t, the most technically sophisticated people in the room will dismiss the rest.

The argument is this: AI does have judgment. When it evaluates five alternative outreach emails and selects the one most likely to generate a response, it is performing something that looks like judgment. It’s weighing variables, comparing alternatives against criteria, and making a selection. When it reviews three pieces of research and identifies which is most relevant, it’s doing something that, if a human did it, we’d call discernment.

And here’s the uncomfortable part: for a significant percentage of routine evaluative tasks, the AI’s selection will be better than what most humans would choose. Not because it “understands,” but because it’s drawing on a vastly larger base of pattern data than any individual carries. The junior rep choosing between five emails is working from maybe two years of experience. The AI is working from patterns across millions of interactions.

For the average case, the AI wins. That’s not debatable.

Where This Breaks Down

But here’s what that argument quietly assumes: that the criteria for evaluation are known and stable.

When the AI selects the “best” outreach email, best according to what? Open rates? Response rates? Conversion rates? Those are measurable, historical, backward-looking metrics. The AI is asking: based on everything that has worked before, which of these is most likely to work again?

That’s not judgment, it’s optimization. And the distinction matters enormously.

Judgment is what you need when the criteria themselves are in question.

When the “best” outreach by every historical metric is wrong for this prospect because they just got burned by a competitor who used exactly that approach. What they need right now is something that breaks every best-practice rule. No data set shows this. No pattern library predicts it. The person who senses it does so because of something the AI fundamentally cannot access: a read on the emotional and political context that exists outside the data.

Or consider research prep. AI can do extensive research in minutes. It can leverage patterns from thousands of past calls and opportunities, access more data than any individual could, and provide highly targeted support for our objectives. But it can’t identify buried assumptions the customer might already have. It can’t determine what’s most important to the customer right now. It can’t know that an hour before the meeting, there was a shift; a reorganization was announced, a budget got cut, a champion lost their influence. The data hasn’t caught up yet. But a person who’s paying attention, who has the smell test, who reads the room, they know.

Three Layers of Evaluation

Not all evaluation is the same. The confusion between three distinct layers is what allows the “AI has judgment” argument to sound convincing.

Layer One: Selection among known alternatives against established criteria. This is what AI does brilliantly. Pick the best email based on historical response data. Choose the most relevant research based on semantic matching. Rank prospects by likelihood to convert. This is optimization, and humans should stop doing it. The AI is better.

Layer Two: Evaluation of whether the alternatives themselves are the right ones. The AI selected the best of five emails, but should you be sending an email at all? Maybe this prospect needs a phone call, or an introduction through a mutual connection, or six months of silence.

AI can only choose among options it’s been given or that it generates from historical patterns. It cannot step outside the frame and ask whether the entire approach is misconceived. This requires contextual judgment and understanding the human situation, not just the data profile.

Layer Three: Evaluation of whether the criteria are right. This is the deepest level. Are we optimizing for the right thing? We’re measuring response rates, but should we be measuring relationship quality? We’re ranking prospects by conversion likelihood, but are these the right customers for where our business needs to go?

This is strategic judgment. It requires not just experience but values, a sense of what matters that isn’t derivable from any data set because it’s about the future, not the past.

AI dominates Layer One. It’s increasingly capable at parts of Layer Two. It is essentially blind at Layer Three. And Layer Three is where the decisions that actually shape individuals and organizations. And they are fundamental to success.

Prisoners of Our Own Making

Everything I’ve described so far, the hollowed-out apprenticeship, the confusion between optimization and judgment, the need to redesign entry-level work around evaluation rather than production; all of it requires something that most organizations don’t have: leaders who can tell the difference.

I’ve written before about what I call “Prisoners of Our Inexperience.” The iconic technology companies that define our industry were built by deeply experienced leaders, people who carried decades of scar tissue from failures, cross-functional moves, and hard-won wisdom.

But the high-growth environments they created systematically produced a generation of leaders with almost none of that experience. People who joined at 22, rode the rocketship, and became VPs by 26, having never worked anywhere else, never sold in a downturn, never rebuilt a failing team, never managed through a crisis they didn’t create.

This is the compounding problem at the heart of the AI transition. The leaders deciding how to integrate AI, which roles to automate, which to preserve, how to develop the next generation, are themselves products of the same de-skilled pipeline. They came up through systems that rewarded execution over inquiry, activity metrics over judgment, template-following over creative problem-solving.

So when they look at entry-level work and see something AI can automate, they automate it. Because from their vantage point, the work was never developing anyone in the first place.

They’re right about the symptom. They’re wrong about the diagnosis. The work wasn’t developing people, but that’s because we’d already broken it, not because it was inherently undevelopable. And the AI decision locks in the damage.

This compounds upward. Mid-level managers who never developed judgment can’t coach their teams toward it. Senior leaders who never developed contextual creativity can’t design organizations that foster it. Each generation builds an environment that produces the next generation in its own image. With the same gaps, the same blind spots, the same reliance on process over discernment.

The Pace of This Transition

At this point, someone usually raises the historical argument: we’ve been through technological transitions before—the web, manufacturing automation, the mechanization of agriculture. And in every case, the economy adapted. New jobs were created. People found new roles. It just took time.

The implication is: relax. This too shall pass.

I’m cautiously optimistic, but I don’t think the historical comparison holds as neatly as people want it to. Three things are different this time.

First, the infrastructure is already in place. When the web arrived, we had to build physical networks, train an entire workforce in basic digital literacy, and create new business models from scratch.

AI is being deployed on top of a mature digital infrastructure. The distribution channel already exists. That removes one of the major historical barriers to adoption speed.

Second, the economic pressure is more acute. Previous transitions occurred when labor markets had greater slack, allowing companies to absorb inefficiencies for longer. Today’s margin pressure creates a stronger pull toward adoption even when the technology is not yet ready. Companies will adopt prematurely, which accelerates both the disruption and the inevitable correction.

Third, AI is recursive in a way that previous technologies weren’t. The web didn’t build more web autonomously. Manufacturing robots didn’t design better robots. But AI systems are already being used to train, evaluate, and improve AI systems. That creates a compounding dynamic that is genuinely different from prior transitions.

I don’t believe the transformation will happen overnight. I’ve seen enough technology hype cycles to know that the gap between promise and delivery is always wider than the enthusiasts claim.

There will be a backlash. There will be high-profile failures: AI-generated legal briefs with fabricated citations, automated customer service that drives customers away, strategies that look flawless on paper and collapse on contact with reality. That practical pain will slow things down.

But I also don’t think we have the twenty-year runway that previous transitions afforded. My estimate is five to eight years of turbulent adoption, correction, and reorganization. Fast enough that individuals and organizations can’t afford to wait it out. Slow enough that the people who invest now in building genuine judgment will have a meaningful window to differentiate themselves.

The Reset

So we’ve made our bed. The question is whether we’re willing to remake it.

I believe we are, because the discomfort is finally impossible to ignore. When a human was producing mediocre, judgment-free work, you could paper over it. The output looked professional. The deck was formatted correctly. The calls got made. It was expensive mediocrity, but it was quiet mediocrity.

In this environment, AI generates mediocrity at scale. When everyone’s using the same tools to produce the same undifferentiated output, the sameness becomes visible. When AI-generated content starts failing in embarrassing ways, it forces a reckoning that decades of human mediocrity never did.

That reckoning is the opportunity.

Redesigning Apprenticeship

The path forward isn’t removing AI from entry-level work. That ship has sailed. The path is redesigning the nature of entry-level work so that AI handles production while humans are tasked with the judgment layer from the beginning.

Instead of “write the report,” the entry-level task becomes: the AI wrote three versions of this report, identify which one the client will actually respond to, and explain why. That forces contextual judgment. It develops the smell test. It builds the pattern library and the professional voice.

Instead of “research these fifty prospects,” the task becomes: the AI has profiled these fifty prospects; find the three where the AI’s assessment is probably wrong, and explain what it missed. That requires understanding human motivation, organizational politics, and the limitations of data. It develops agency because the person has to commit to a contrarian position and defend it.

Instead of “make the calls and follow the script,” the task becomes: the AI generated a call script for this prospect, rewrite the opening based on what you know about this person that isn’t in the CRM. That preserves the beneficial friction of constraint-based learning while pushing creative muscles earlier.

The common thread: the human’s job shifts from producing the output to evaluating and improving it through contextual knowledge the AI doesn’t have. That’s creativity. It’s also agency, because every evaluation is a choice, and every choice carries the risk of being wrong.

This is harder work than what most entry-level people do today. It’s also more interesting, more developmental, and more valuable. But it requires managers who can coach judgment rather than just monitor activity metrics.

Redesigning Leadership

And this is where the real transformation has to happen.

Everything we’ve discussed: redesigning apprenticeship, redefining entry-level roles, building organizations around judgment rather than production, —requires leaders who themselves possess the judgment, creativity, and contextual awareness to design and coach these new systems.

That means leaders have to do something uncomfortable: acknowledge that the system they came up through didn’t develop them the way they thought it did. The AI is showing them that. And instead of defending the system or automating it further, they have to use this moment to build something better.

This starts with three shifts.

From measuring output to developing capability. If the deliverables are now AI-generated, what does a manager measure? Not volume. Not speed. The measure becomes: can this person identify what the AI got wrong? Can they explain why? Can they improve it in ways that reflect genuine understanding of the customer, the market, the situation? Evaluating these things requires managers who have those capabilities themselves.

From providing answers to designing friction. The instinct of most managers is to make things easier for their teams. Remove obstacles. Streamline processes. But in a world where AI has already made production effortless, the manager’s job is to reintroduce the right kind of friction. Not busywork, but friction that forces people to think beyond the first answer, to question assumptions, to develop the judgment that only comes from struggling with complexity.

From defending experience to seeking it. The “Prisoners of Inexperience” problem doesn’t fix itself. Leaders who recognize the gap in their own development and have the humility to address it will build very different organizations than leaders who believe their trajectory up the rocketship has taught them everything they need.

Seeking cross-functional exposure, engaging with perspectives outside your industry, studying how problems were solved in contexts radically different from your own—these are the antidotes to the narrowness that the growth machine produced.

The 80%

Everything I’ve described so far, the redesigned apprenticeship, the leadership shifts, the move from production to judgment, reads like advice for the top 20%. The people who already seek friction, already question the playbook, already develop judgment through sheer force of curiosity. Telling them to do more of what they’re already doing isn’t particularly useful.

The real question is: what about everyone else?

In any organization, performance follows a distribution. The top 20% are the ones who’ve always operated with contextual judgment, whether the system asked them to or not. The bottom of the distribution will always struggle. But the 80% in the middle, is where this article either becomes a provocation that changes behavior or a think piece that people nod at and forget.

An Honest Look at the 80%

Let’s be direct about who they are, without condescension. The 80% aren’t lazy or lacking intelligence. They’re rational actors responding to the system they’re in. For decades, the system rewarded them for exactly what we’re now saying is insufficient: executing the process, hitting the metrics, following the playbook.

They optimized for what the organization measured, and the organization measured output, not judgment. They did exactly what they were told success looked like.

So when we say “develop contextual judgment, exercise agency, embrace friction,” many of them hear something that sounds like “be a fundamentally different person than the one this organization hired, trained, and promoted.” That’s not a small ask. And without a clear picture of what this looks like at their level, it’s an impossible one.

The 80% isn’t monolithic. When considering a distribution shift, it is helpful to identify at least three distinct groups, each with different barriers and leverage points.

The willing but untrained. Maybe a quarter of the 80%. These are people who sense that something is changing, who feel the discomfort, who might even be reading articles like this one.

Their barrier isn’t motivation; it’s that nobody has shown them what “exercising judgment” looks like in their specific role. They’ve been in production mode so long that they don’t know how to shift.

They need models, examples, and permission. For this group, the redesigned apprenticeship I described earlier isn’t just for new hires. It’s development for mid-career professionals who were never developed the first time around.

The comfortable middle. Probably the largest group. They’re performing adequately by current metrics. Their jobs haven’t been automated yet. The urgency doesn’t feel real. They’ll change when the pain of not changing exceeds the pain of changing, and not a moment before.

For this group, the question isn’t “how do we inspire them. It’s “how do we change the system so that the thing being measured and rewarded is judgment, not just output.” They’ll follow the incentives wherever the incentives point. If the incentives still reward activity volume and process compliance, that’s what they’ll deliver, regardless of how many articles they read about the future of work.

The actively resistant. They’ve built their identity and their career around being excellent executors. Telling them that execution is being commoditized isn’t just a professional threat, it’s an identity threat. They’ll push back, dismiss the argument, double down on what’s worked. Some will come around when the evidence becomes undeniable. Some won’t, and organizations will need to make hard decisions about that.

What the 80% Can Do

For the willing but untrained, the path is clear and immediate. Start with the work you’re already doing and add one layer of judgment to it. Not a transformation, an addition.

You ran the AI-generated analysis? Before you do it, think about what you are really trying to achieve. It’s not hitting the activity numbers, but it’s the outputs. Focus the AI analysis on that.

Then, when AI produces the results, before you send it, go back to your original question. Does it produce what you needed it to produce? What do you have to change to have it work more effectively? Do you have to go back and reassess what you asked?

These are small acts of contextual judgment. They take five minutes. They don’t require anyone’s permission. And over time, they build exactly the critical muscles, the smell test, the pattern library, the professional voice.

The key is that they’re additive to the current work, not a replacement. You’re not being asked to stop doing what you know how to do. You’re being asked to notice what you think about what you’re doing.

For the comfortable middle, the lever is organizational, not individual. This is where the leadership argument becomes critical. The 80% won’t shift until the system shifts. That means managers have to start asking different questions. Not “did you complete the task?” but “what did the AI miss?” Not “how many calls did you make?” but “what did you learn that changed your approach?”

Those questions change what gets rewarded. And what gets rewarded changes behavior. The 80% doesn’t need to be inspired. They need to be redirected by leaders who have the courage to measure what actually matters.

What Happens If They Don’t

The distribution will shift whether individuals choose to shift with it or not. AI is compressing the bottom of the value curve. Work that requires only Layer One evaluation, selection among known alternatives, is heading toward zero economic value. Not because the work disappears overnight, but because AI can do it for effectively nothing. The human represent only cost

The 80% who don’t develop Layer Two and Layer Three capabilities don’t get fired in a dramatic AI-replaces-everyone moment. What happens is quieter and in some ways worse: they become interchangeable.

Any one of them can be replaced by any other, or by a less experienced person with better AI tools, or eventually by the AI itself.

Their careers flatten. Their compensation stagnates. Their professional identity erodes not through a single event but through the slow realization that the market no longer values what they do.

The organizations that don’t shift face the same compression. They become commodity providers. Their output looks like everyone else’s because it’s generated by the same tools following the same optimization patterns.

They compete on price because they can’t compete on insight. And competing on price against AI-augmented competitors is a race to the bottom that humans can’t win.

Shifting the Distribution

We probably won’t change the shape of the normal distribution. There will always be a top 20% and a bottom 20%. But the question isn’t whether we can eliminate the curve, it’s whether we can move the whole thing to the right so that the median professional is operating at a level that today might represent the 65th or 70th percentile.

The mechanism for that shift is exactly what we’ve been building toward: redesign the system, not the individual. Change what entry-level roles develop. Change what managers measure and coach. Change what leaders model and reward. Create environments where beneficial friction is the norm rather than the exception. Make the exercise of judgment a daily practice rather than a rare event reserved for senior people.

The top 20% will still outperform. But if the median performer is exercising contextual judgment rather than just executing processes, the entire organization operates at a fundamentally different level.

The gap between “average” and “excellent” narrows in terms of baseline capability, even as it persists in terms of the transformational leaps that only the best will make.

That’s the shift. Not turning everyone into a visionary. Just raising the floor from “can follow a process” to “can evaluate whether the process is right.” That alone would be revolutionary for most organizations.

The Human Margin

Let me come back to where we started: the question of consciousness.

We’re asking whether AI is conscious because we’re trying to locate the boundary between what machines can do and what remains uniquely, irreducibly human. And I think the answer is simpler and more challenging than most of the headlines suggest.

AI can optimize. It can select the best option from a known set against established criteria, faster and more consistently than any human. For the vast landscape of combinatorial, production-oriented work, the work we’ve been doing on autopilot for decades, it’s not just adequate. It’s superior.

But AI cannot originate purpose. It cannot decide what’s worth doing. It cannot feel the tension between how the world is and how it should be. It cannot bear the weight of being wrong about something that matters. It has no skin in the game.

The human margin, the part of us that no amount of parameter scaling can replicate, lives in the friction. In the struggle to understand a specific human being in a specific moment.

In the willingness to throw away the playbook when the situation demands something the data has never seen. In the judgment that comes not from processing more information, but from caring about the outcome.

For decades, we minimized this. We built organizations that treated it as a distraction rather than reality. We promoted people for predictability and penalized them for the kind of independent thinking that makes organizations genuinely adaptive. We made our bed.

AI is the wake-up call. Not because the machines are coming for our jobs, but because they’re showing us how much of what we called “work” was never the point.

The point was always the judgment, the creativity, the agency, the consciousness that we were supposed to be developing in ourselves and in the people around us.

The question isn’t whether AI can think. The question is whether we’re willing to start thinking again, really thinking, with all the friction and discomfort and risk that implies.

The machine has the map. But we’re the only ones who know where we actually need to go.

And that’s not a burden. It’s a privilege.

Afterword: This is a fascinating AI generated discussion about this post. They capture the ideas very well. Enjoy!

Recently, I was listening to a podcast. It’s one of my favorites, The Transaction, hosted by Craig Rosenberg. The discussion was outstanding, but inevitably, it shifted to discussing a “new and novel” practice that was game-changing.

In this case, the discussion was on demos. The idea is, “We discovered something revolutionary. Rather than our standard demo, we demo using the customer’s data!” It’s an outstanding practice; it drives higher engagement and relevance with the customer.

But here’s my problem. When I was going through my training at IBM (sometime in the last century, but I won’t tell you which half), this was standard practice. You always demo’d using customer data; you never do a standard demo. And sometimes, we sat customers behind the keyboard (yes we had them in those days) and guided them through demo-ing it for themselves.

I offered the observation to Craig, “What would happen if we could stop reinventing the wheel, but study and adapt great things from the past or from other industries? Wouldn’t we reduce the time and risk to results? Wouldn’t we learn and adapt faster?”

It’s a question I keep coming back to because I run into it all the time, particularly in modern tech industries, including SaaS and new AI-native startups. And the more I look at it, the more I see a pattern that goes far deeper than any single practice.

Iconic companies that defined SaaS enterprise software today, Salesforce, Workday, ServiceNow, and others, were all built by deeply experienced leaders. People who had spent decades in the trenches, who carried scar tissue from failures and hard-won wisdom from navigating complex organizations. They didn’t stumble into building great companies. They built them because of everything they’d learned.

But the SaaS industry they spawned? Those new companies have systematically produced a generation of leaders who have almost none of that experience.

And we’re paying the price.

The Experience Gap

Let’s look at how this happened.

Many of these startups were fueled by cheap capital and explosive segment growth. People with great product ideas got seed funding to build a great technology. They may have had past experience, but usually it was in product or operational roles. And increasingly, many had virtually no experience at all, just an idea that could capture funding in a cheap capital market.

These organizations created an environment where someone could join at 22 and become a VP by 26. Without a doubt, that progression was the result of outstanding performance. But the rapid expansion meant there were always more roles than qualified people to fill them, so promotion became a function of timing and availability, perhaps ignoring readiness.

This isn’t a criticism of these individuals. They’re smart, driven, and capable. The problem is the underlying operating model. It is optimized for growth velocity rather than simultaneous leadership development. It confused scaling with capability. And it created leaders whose entire frame of reference is a single operating model, single company, a single market, and a single economic context.

Think about what that means in practice.

These leaders have never sold in a sustained downturn. They’ve never managed through a restructuring. They’ve never worked in an industry where the rules were different, where customers behaved differently, where the competitive dynamics forced entirely different strategic thinking.

They operate, in effect, with a single set of experiences, and they treat those as universal truths. And in SaaS, these sets of experiences were reinforced across thousands of companies. Everyone saw things the same way. Everyone had the same growth experience. People moved from company to company, always applying the same playbooks.

As each organization looked at others, the confidence in this operating model was reinforced.

So when they encounter problems, too often they’re seeing those problems for the first time. They have no one on their teams who’s experienced them before. They’ve never studied how other companies solved them.

Watch closely and you’ll see companies A/B testing their way to insights that an experienced leader would have brought on day one. You’ll see go-to-market strategies being “innovated” that are actually just rediscoveries of principles that experienced sales and marketing leaders have known for decades. You’ll see organizational structures being experimented with and failing in predictable ways.

Despite these problems being very old and well known, because of their inexperience, these problems are new to them.

This is expensive learning. Expensive in time, risk, results, and in the careers of the people caught in the churn of constant reorganization.

And there’s a cultural dimension that makes it worse. In many of these environments, “we’re different” has become an identity. Outside experience gets discounted rather than valued. “That’s how they do it at legacy companies” has become a dismissal rather than a learning opportunity. The insularity that limits these leaders’ perspective also prevents them from recognizing the limitation.

For someone like me, looking at this from the outside, I tend to react, “I’ve seen this so many times before. I made the same mistakes 30 years ago. I’ve seen so many others making the same mistakes and coming up with solutions over the decades.”

What troubles me is that these well-intentioned, high-performing individuals don’t realize others have been there and there are solutions that can be updated and adapted very quickly. They are condemned to reinventing wheels that have been around for decades.

This Isn’t About Blame, It’s About Recognition

I want to be clear about something. This is not an Old Fart whining about “kids these days.” It’s not about age, and it’s not about talent. Some of the smartest, most energetic leaders I encounter are in these companies.

The problem is systemic. These organizations never built the leadership development infrastructure that matches the complexity they now face. They promoted people quickly without giving them the breadth of experience that builds genuine strategic judgment.

They created cultures that celebrate internal knowledge over external perspective. And they did all of this during a historically unusual period of low interest rates and abundant capital, a period that rewarded growth above all else and papered over a lot of strategic mistakes.

Now the context has shifted. Growth has slowed. Capital is expensive. Markets are more competitive. The playbooks that worked from 2015 to 2021 don’t work anymore. And that’s when inexperience becomes most dangerous. The context has changed, and most have never navigated this before.

The Dunning-Kruger dimension is real here. When you’ve only ever operated in one context and it worked, you don’t know what you don’t know. Your successes have taught you less than you think, because you can’t separate what worked because of your decisions from what worked because the market was lifting all boats.

Gaining the Experience You Don’t Have

So where does that leave us? If you’re a leader who recognizes this gap, and that recognition itself is the critical first step, what do you do about it?

Inexperience is a condition, not a character flaw. And it’s one you can actively address.

Spend real time with customers. Not reviewing dashboards. Not conducting “customer visits” that are really product pitches in disguise. This is genuine, sustained engagement with customers: sitting in on their meetings, understanding their workflows, hearing about their struggles in their own language.

When you do this well, you’re borrowing their experience. You’re learning what they learned the hard way about their own business, their own industry, their own challenges. And many of these companies may be your customers, they would welcome thoughtful conversations about what they have learned. This is the most accessible lever any leader has, and it’s chronically underutilized.

Study outside your industry; not casually, as a discipline. How did companies navigate the transition from growth to maturity? What happened to enterprise hardware companies in the 2000s? How do professional services firms develop partners? How did P&G build the brand management system that entire industries adopted?

The SaaS leader who studies these questions will see patterns that their peers completely miss. The challenges of scaling organizations, motivating people, navigating market shifts, managing through downturns, none of this is new. There is a deep body of knowledge available to anyone willing to look beyond their own industry’s boundaries.

Seek mentorship from outside your ecosystem. Not an advisory board of other SaaS executives who all share the same blind spots. Genuine relationships with leaders from different industries, different eras, different contexts. The value isn’t that they give you answers. The value is that they ask you questions you’ve never considered, questions that expose assumptions you didn’t know you were making.

Push for deliberate rotation and stretch assignments, for yourself and your team. This has been one of the most powerful experiences in my personal development. While I started in sales, as I grew in the organization, I was moved into different functions: product, strategy, manufacturing, and development. They forced me to learn and develop new perspectives. While I kept coming back to more senior sales roles, I had a more diverse experience base to leverage.

The “old school” companies did this well. Leaders didn’t go from rep to manager to VP in one straight line at one company. They moved across functions, across geographies, across customer segments. Each move forced them to be a beginner again, and that discomfort was the learning.

Modern tech companies could do this, but rarely do, because they think it slows scaling. But scaling has hit a roadblock and organizations are being forced to change. But they don’t have the experience base to know what to change and how.

The Prerequisite

Everything I’ve described above requires one thing above all else: humility.

The humility to recognize that your success may have taught you less than you think. The humility to seek out perspectives that challenge your assumptions rather than confirm them. The humility to admit that what you know, what made you successful, may not be helpful in continuing to grow and succeed. The humility to look at your failures, and every leader has them. It’s the humility to ask, “Is there a better way?” rather than assuming the market simply wasn’t ready.

Our inexperience is not an excuse. It’s a fact. But we can’t let that inexperience limit us. If we do, we will continue to be prisoners of our inexperience.

Afterword: This is an interesting AI narrated discussion of this article. There are a few small errors, they make some arguments I don’t totally buy, but the perspective is interesting.

We have learned that greatness comes from focused effort: 10,000 hours of experience, clear vision, steady progress. While these are important, the reality is messier.

Startups don’t have the luxury of neat plans and focus. Founders perfect products while simultaneously selling, recruiting, managing cash flow, and fixing the printer. There’s no blueprint. Every decision is made with incomplete information: Do we pivot or persist? Is this signal or noise? Right hires? Right pricing? Right market?

Most don’t fail from one catastrophic failure. They die from attrition. The customer who doesn’t renew, the key employee who leaves, the feature that ships late, the new technology that turns development upside down, the continued challenge of maintaining funding. Survival means absorbing these setbacks while projecting confidence you don’t feel.

For founding teams, this struggle creates its own energy. Every setback is data. You have nothing to lose. The target is clear: survive today, traction tomorrow, scale eventually. This hunger is the startup’s secret weapon. As you start to succeed, a founder’s mindset develops, built on the collective experience of achieving greatness.

When you finally emerge as an established company, you’ve earned something real. You’ve been tested. Not only survived but achieved success.

What we miss is that the moment we achieve success, everything is about to change.

The Inversion

The moment success is recognized, the rules transform. Unconsciously, you start to move from offense to defense. From disruptor to disrupted. From hunter to hunted.

As a startup, every risk carried lopsided upside. Failure meant returning to zero; success meant breakthrough. Now things are inverted. You have something to lose. What previously looked like a risky but bold bet is now viewed as risky and foolish.

Competitors study your success as their roadmap. Your innovations become their baseline. You’re running faster to stay in place. That 20% growth from $10M to $12M looks different when you’re at $100M.

Your failures, once only known internally, are now public and amplified. Pressure builds toward safety, predictability, protecting the brand. The risk tolerance that made you great is displaced by cautiousness.

The habits that built success: early mornings and long days, obsessive detail, paranoid customer responsiveness begin to soften. Newer employees lack the mindset critical in your early growth. Your own days fill with internal meetings, reporting, maintenance, and routine.

The discipline erodes. The accountability diffuses. You become prisoners of the methods that worked before even while they are not working as well as they originally did.

The Next Peak

Here’s what separates the temporarily successful from the enduringly great: the journey doesn’t end with this first summit. When you reach it, higher peaks become visible. Your ambitions to grow, to reach those peaks continue. But the path doesn’t lead up from where you currently stand. You won’t reach Peak B by continuing everything you have always done.

This is where we get scaling wrong. We think the formulas that worked to reach the first peak can be continued to reach successive peaks. But look at Salesforce, Microsoft, Nvidia, Google. As each moved from initial success, they continued to transform, developing strategies, capabilities, people, processes, tools, products critical to reach Peak B, then Peak C, then Peak D.

Startups by their nature start in the flatlands. They identify a peak to conquer. And in this process, they know they will either survive or die. The startup’s first chasm is brutal, but it has one advantage: you’re small, hungry, and unencumbered. You have nothing to lose.

We often think moving to the next peak is voluntary. We could be satisfied with current success, choosing to stay, maintaining and slightly growing revenue.

But those pursuing greatness are unsatisfied. They see the next peak, the next challenge, the opportunity to grow in ways that are different and better.

Why the Second Crossing Is Harder

Moving to this next peak, everything is different. The things that worked for you now work against you.

Organizational mass. A 50-person startup pivots in weeks. A 5,000-person company takes years. Every layer of management, every process optimized for current business, every system that “works” now become a drag. The resistance isn’t malicious. It’s the nature of transforming large organizations.

Infrastructure burden. You’ve built an ecosystem: suppliers, partners, integrations, customers. Complex organizations focused on their own priorities, massive data requiring integration, planning and funding processes, people challenges, metrics and dashboards tracking everything. The startup attacking you only needs to create a single winning product. You’re carrying the weight of everything your prior success created.

Scaling math. Our perspective on scaling changes at that first peak. As early startups, 2x growth, moving from $5M to $10M, is achievable, particularly in a hot space. Achieving 2x at $100M or $1B is hugely different, probably impossible. High growth expectations adjust. It may be 10%, taking the $100M company to $110M. The lens through which we measure success must change to sustain greatness.

Customer expectations. Your early adopters tolerated incomplete products. They bought the vision and wanted to be part of it. Current customers are pragmatists. They signed contracts expecting to realize the value committed. They expect stability, predictability, the complete solution you promised. You can’t abandon them, but serving them anchors you to Peak A.

Revenue dependency. More people depend on current revenue streams: employees, shareholders, partners, suppliers. The startup attacking you can burn cash for years chasing the future. You have quarterly earnings calls and boards expecting predictable, profitable performance.

Talent mismatch. You hired people to scale Peak A, those who excelled in experimentation, nimbleness, agility. But skills shifted at that peak. Your needs moved to operators skilled at protecting, defending, maintaining success. Efficient execution of standard processes.

Moving from Peak A to B needs people with mindsets similar to your startup, but with a difference. People comfortable with ambiguity, willing to break what works. These people don’t want bureaucracy, and operators don’t want chaos. Both must co-exist in the same organization. Maintaining standards while driving growth.

Cultural calcification. “This is how we won” becomes “this is how we do things” becomes “this is who we are.” Success breeds convention. The longer you’ve been successful, the deeper the identity investment in current methods. Resistance to change isn’t just institutional, it’s personal.

Governance constraints. Startups answer to a handful of investors who understand long-term bets. Public companies face analysts who punish margin compression, activists who demand short-term returns, boards and customers that expect predictability. The structure itself resists transformation.

The startup crossed the first chasm with nothing to lose. The established company must cross the second while protecting what it built. That instinct is what makes the next stage so difficult.

Those Who Couldn’t Cross

History is littered with great performers that couldn’t make the transition.

Kodak invented the digital camera in 1975. They saw Peak B with perfect clarity. But descending from film meant cannibalizing their most profitable business, alienating customers who wanted film products, dismantling the ecosystem of developers, retailers, and paper suppliers. They had the vision. They lacked the courage, or perhaps the structural freedom, to make the transition.

Blockbuster had the chance to buy Netflix for $50 million. But late fees generated $800 million annually. Their customers, families who wanted convenient weekend rentals, weren’t asking for mail-order DVDs. The whole product was optimized for a certain business model. They defended this position in a market the world was abandoning.

Nokia dominated mobile phones. BlackBerry owned enterprise. Both saw smartphones coming. Both had resources and talent to respond. Neither could sacrifice current profits, current customers, current organizational identity for uncertain transformation.

Intel drove the PC revolution, yet was unable to transition into modern AI semiconductor technologies.

Those Who Made the Crossing

Other companies made that transition, often completely transforming their businesses.

When I worked for IBM, mainframes dominated but we started to be threatened by mini and micro computers. In the early 1990s the mainframe business was collapsing, $8 billion in losses, 60% stock decline. Lou Gerstner led a brutal transformation: massive layoffs, exiting commodity hardware, killing sacred product lines. He rebuilt IBM around services and consulting. The company that sold boxes learned to sell solutions. IBM learned to transform again, pivoting to cloud and AI.

Microsoft under Ballmer clung to Windows and Office while mobile passed them by. Satya Nadella took the company down from that peak, embracing Linux (Ballmer once called it “a cancer”), open-sourcing core technologies, betting everything on cloud. Azure cratered Windows licensing. Longtime customers were confused. The organizational identity crisis was real. But Nadella demonstrated the curiosity to see the new peak and the discipline to march toward it. Azure is now neck-and-neck with AWS. Microsoft is one of the new leaders in AI.

Netflix killed their own DVD business, a profitable operation with loyal customers, to bet on streaming. Adobe abandoned boxed software for subscriptions, watching revenue plummet for quarters before it soared. Apple cannibalized the iPod with the iPhone, knowing it would destroy a product line that dominated its category.

Each descended voluntarily. Each endured confusion, criticism, metrics decline. Each required leaders with resilience to absorb organizational pain and accountability to own decisions that looked wrong before they looked right.

What the Crossing Demands

Making the transformation requires specific mindsets and behaviors that must be cultivated before they’re needed:

Curiosity and continued learning. We have to have the curiosity to search for Peak B when you’re comfortable on Peak A. Leaders who sustain greatness maintain genuine interest in what’s changing, what’s emerging, what could disrupt them. They seek dis-confirming information. They stay students.

Customer focus. An obsession with understanding customers and markets. Not just where they are now, but where they are going. Understanding what will impact them in the future.

Discipline to maintain standards during the descent. When metrics decline and confidence wavers, the temptation is to cut corners, lower bars, accept “good enough.” The crossing demands holding standards tighter, not looser.

Accountability to own decisions that won’t validate for years. The leader who initiates the descent may not be around when the new peak is reached. Crossing requires people willing to be accountable for outcomes they may never see.

Ability to deal with change and complexity. Moving from Peak A to B is filled with unknowns and complexity. It demands resilience to absorb the organizational trauma of transformation. People will leave. Metrics will suffer. Critics will multiply. The crossing isn’t a sprint, it’s a sustained commitment to achieve a goal and not to be stuck in the comfort of where you currently are.

Caring enough about the organization’s future to endure the pain of its present. Transformation is hard on people. Leaders who sustain greatness bring their people through it—not around it, through it.

Purpose. A relentless focus on what you want to stand for—for customers, people, partners, suppliers, shareholders. A “true north” focused on achieving your full potential.

The Audit

If you’ve achieved some measure of success, ask yourself:

On Risk: Am I making decisions to gain something new, or to avoid losing what I have?

On Learning: When was the last time I was genuinely a beginner at something relevant to my field?

On Standards: Have I lowered my minimum acceptable standard because I’ve “earned” a break?

On Growth: Can I see Peak B? What’s really stopping me from making the crossing?

The Day One Test: If I lost everything and had to rebuild, would I use the same strategies I’m using now? If not, why am I still using them?

The Point

Achieving greatness is an event. Sustaining it is a system.

The leaders who endure aren’t those who climb highest the first time. They’re those willing to descend and climb again—carrying an organization that resists, serving customers who didn’t ask for change, dismantling systems that still work.

The summit isn’t the point. The climbing is the point. The willingness to keep climbing, even when everything you’ve built makes the next climb harder.

The question isn’t whether you can reach the peak. It’s what you’ll do when you see the next one.

Afterword: Another outstanding AI generated discussion of this article. Enjoy!