He calls a meeting. The right people are in the room. And somewhere in the first twenty minutes, someone names it.

The problem is silos. We're not working across functions. Everyone's focused on their own lane.

The room nods. Because the room knows. Everyone in that room has felt the friction, lost time to the handoffs, wondered why information that existed somewhere in the organization never seemed to reach them.

From somewhere near the end of the table, someone says it quietly.

We've talked about this before. I remember we did something about silos — maybe ten, twelve years ago.

A beat.

What happened? someone asks.

I'm not sure exactly. It kind of faded out.

A brief pause. Then the room moves forward, because there's a problem to solve and a plan to make. And this time, everyone agrees, they're going to do it right.

I know that room well. A few decades ago I was James — the one with the energy, the idea, the certainty that this time we'd get it right. Now I'm the frustrated one at the end of the table, saying we tried something like this in the '80s, watching the room absorb it and move on anyway.

Both of us are right. Neither of us has the whole picture.

—

They make a real plan. Not a vague commitment — an actual structure. A cross-functional team, pulling one person from each key function. A standing meeting, every two weeks. A shared mandate: bring your stuck problems here, look across the boundaries, solve together.

People leave the room with energy. James feels it. This is how it should work.

—

Fast forward a year.

The cross-functional team has stopped meeting. It didn't end dramatically — there was no moment of dissolution, no announcement. The meetings just got harder to schedule. People were pulled into other priorities. The standing agenda got thin. Eventually someone suggested they reconvene when there was something specific to bring, and that moment never quite arrived.

The silos are still there. The handoffs still drop work. Decisions still take too long. And here's what's strange: the conversation about silos has disappeared too. Not because the problem was solved. Because the energy around solving it was spent, and the organization moved on to the next thing.

If you asked James today, he'd probably say they tried something and it didn't really take. He might not remember exactly why.

In ten years, he'll be the one at the end of the table.

—

There's a simulation used in business schools called the Beer Game. Players take on roles in a supply chain — retailer, wholesaler, distributor, manufacturer. Their goal is simple: keep the beer flowing without running out or overordering.

They always fail. Not because the players are careless or inexperienced. Because the structure of the game — the delays, the information gaps, the incentives baked into each role — produces the failure regardless of who's playing. You can swap in the smartest people in the room. The outcome doesn't change.

The system determines the behavior. Not the people inside it.

That's what happened to James's cross-functional team. Every person on it went back to a boss who evaluated them on their function's results. A budget that belonged to their department. A performance review that asked what they delivered for their team — not what they did for the whole. The team asked people to think horizontally. The structure around them rewarded vertical. Collaboration was the stated goal. The path of least resistance ran the other way.

The team didn't fail because people stopped caring. It failed because the structure never changed. And structure, left untouched, keeps producing the same outcome — regardless of the plan made in the meeting room, regardless of the energy people walked out with.

The silos didn't come back after the team disbanded. They never went anywhere.

—

Which raises the harder question: if the structure is the problem, why doesn't anyone change it?

Because structure isn't just a set of org chart boxes. Every organization carries what I think of as cultural DNA — the invisible instructions that define the organization. It lives in budgets, incentives, reporting relationships, performance measures, and countless decisions made years ago that no one questions anymore.

The organizational structure creates the conditions. The human nervous system adapts to those conditions. Over time the two reinforce each other until what feels like common sense is simply the encoding made visible.

Nobody in James's organization chose silos. The structure made silos the easiest, safest, most rewarded behavior available. People followed the logic. Of course they did.

—

This is why I keep saying it's the '80s again.

I've sat in versions of this meeting since the 1980s. The slides change. The language changes. The consulting firms change. But eventually someone says we're too siloed, someone forms a cross-functional team, and everyone leaves convinced this time will be different.

It isn't that organizations can't learn. It's that learning without redesigning the structure is like pulling weeds without touching the root. The symptom disappears. The system keeps growing it back.

New leaders inherit old systems. Old systems keep producing familiar outcomes. The problems look new because the people experiencing them are new. The structure producing them is not.

Different technology. Different corporate speak. Different people. Same outcome.

—

The Beer Game doesn't change when smarter players sit at the table. It changes when the rules change. Organizations are no different.

The work that actually changes things doesn't begin with what the organization is doing wrong. It begins with what the system is making easy, safe, and rewarded — and what it's making costly, risky, and invisible.

Until an organization changes what it measures, funds, and expects, people aren't choosing silos. They're simply following the logic the structure encoded long before they arrived.

That's where the redesign begins. Not in the meeting room. In the architecture beneath it.

Next: The Roles Grow Back — why organizational structure always rebuilds itself, and what's really holding it in place.

Structure teaches culture what to become.

Have you ever opened the refrigerator looking for the ketchup and not been able to find it? You stare directly at the shelf where it should be. Move things around. Look again.

Nothing.

Then someone walks over, points to the bottle sitting inches from your face, and suddenly it’s obvious.

There it is.

What’s interesting isn’t that the ketchup was hidden. It wasn’t. Your brain filtered it out. We’ve all done it — looking for our keys while holding them, searching for our glasses while they’re sitting on top of our head, reading the same sentence three times and still missing the typo.

The brain is constantly filtering. It has to; there’s simply too much to notice everything. Most of the time that filtering serves us. Occasionally it hides something that’s been in front of us the whole time.

It makes me wonder whether organizations have their own version of the ketchup bottle problem.

We’ve spent decades trying to get transformation right — methodologies, governance structures, operating models, reporting processes, maturity assessments, certifications, entire industries built to help organizations change. And yet transformation remains one of the most frustrating challenges organizations face.

That raises an uncomfortable question. What if the problem isn’t that we haven’t looked hard enough? What if there’s something sitting right in front of us that we still haven’t learned to see?

Here’s the idea I keep coming back to. The most important things organizations miss are rarely hidden. They’re sitting in plain sight, held in place by assumptions so familiar we’ve stopped noticing them. We don’t fail for lack of information. We fail because our assumptions quietly decide what counts as information in the first place.

The default approach

Most transformation work begins with an assumption so familiar we rarely question it: that transformation is fundamentally an execution challenge. Define the future state. Build the roadmap. Establish milestones. Execute the plan. The better we execute, the more likely we are to succeed.

It sounds reasonable — until we remember the world doesn’t stand still while we’re executing. Markets change. Technologies evolve. Leaders leave. Customer expectations shift. A roadmap that was perfectly reasonable when it was drawn can drift out of step with reality while we’re still following it.

And yet we often keep measuring success by how closely we followed the original plan. At some point, discipline stops keeping us on course and starts keeping us from noticing the course has changed.

The climbing-a-mountain approach

Standing at the base, you choose the best route you can see. Halfway up, something becomes visible that was impossible to see from below. A washed-out trail. A safer path. A hidden valley. A better destination.

Was the climb wasted? Of course not. The climb created the information — it only became available because you climbed. Progress and learning weren’t separate activities. They were the same activity.

So why do so many organizations still treat them as separate — learning first, execution after? Reality is messier than that. In a changing environment, progress creates information, information changes decisions, and decisions change the path. Adaptation isn’t a deviation from the plan. It’s often evidence that we’re paying attention.

Why pointing isn’t enough

Come back to the ketchup bottle for a moment. In the refrigerator, the bottle is passive — the moment someone points, the spell breaks, and it stays broken. Organizational blindness doesn’t work that way.

What an organization fails to see isn’t sitting quietly on a shelf. It’s being held in place — by every process, every incentive, every reporting line, every quiet assumption about how things are done here. So when someone names the obvious problem, the organization can nod, agree, and carry on exactly as before. The insight didn’t fail because it was wrong. It failed because the thing it described was held in place by too many hands to move with a single sentence.

Which is why seeing was never the hard part. Sustaining sight — against a system built to restore the old view — is.

The Transformation Office approach

Today, many organizations are creating Transformation Offices. The interesting question isn’t whether that’s a good idea. It’s what problem those offices are being asked to solve.

If transformation is primarily an execution challenge, then reporting, governance, and oversight make perfect sense. If it’s fundamentally an adaptation challenge, then something else is needed — and it’s harder than pointing. The greatest value isn’t tracking what happened, or even helping leaders see what they currently cannot. It’s helping the organization keep seeing: holding a clearer view in place long enough that it doesn’t quietly reassemble into the old one.

In other words, helping the organization avoid its own ketchup bottle problem — not once, but for as long as it matters.

We’ve all had the experience of staring directly at something and not seeing it until someone pointed it out. None of us can name our own ketchup bottle — that’s the whole problem. So the better question is: who in your world is positioned to see what you can’t? And are you actually listening to them?

This is the most ordinary thing a human being does. We are, by a wide margin, the most adaptive organism on the planet, and most of that adapting happens without strain, without a plan, and without our noticing it’s underway. Hold that fact in one hand.

In the other, hold this one: organizations spend enormous sums trying to produce exactly this kind of change in their people, and mostly fail. The whole subject of this essay lives in the space between those two hands — between how effortlessly people adapt on their own and how reliably they resist when an organization tries to make them. What follows is about organizations, because that’s where the gap is most expensive. But the principle underneath it isn’t really about organizations at all. People are always adapting. The only question worth asking is what they’re adapting to.

I should say early that I wrote about some of this once before, years ago, and got the mechanism wrong. I reached for a tidy bit of brain science that turned out to be more slogan than fact. The shape of the idea held; the explanation didn’t. I’ve come to think that’s fitting, because the thing this essay is about — the slow, unforced revision of a model you didn’t know you were holding — is exactly what happened to my own thinking in the years between. So take this less as a correction issued from authority than as a worked example of the phenomenon it describes.

Not all change is the same change

Here is a distinction most change efforts never make, and it explains a great deal of their failure.

Some changes leave you no choice. The payroll system is being replaced; on Monday, the old one is gone, and you enter your hours in the new one. You might grumble, but you’ll learn it, because there is no alternative and not much at stake. The benefit is obvious or irrelevant, the path is single, and your sense of who you are is nowhere near it. People adapt to these changes the way they adapt to a detour around roadwork. There’s no mystery here and, frankly, no need for anything elaborate. One of the quieter truths of this field is that organizations massively over-manage these simple changes — wrapping a routine system swap in engagement plans and resistance-management when people would have just adjusted.

Then there is the other kind. The kind where choice quietly remains. Where a person can comply on the surface and continue exactly as before underneath — nod in the meeting, attend the training, and then go back to their desk and do what they’ve always done. Where the change asks them not merely to do something different but to think differently, to become a somewhat different person at work. Leadership transformation. Building a culture of experimentation or innovation. Working alongside artificial intelligence. These are the changes organizations are actually struggling with right now, and they share three features that the payroll swap does not.

The first is that choice survives. There is almost always a way to route around the new behavior while appearing to adopt it.

The second is that a mental model has to move. You can’t get there by learning a procedure, because the obstacle isn’t a skill gap — it’s a way of seeing.

And the third is the one that matters most, and the one almost everyone underestimates: the old way still works. The leader who rose through command, expertise, and control is not being asked to swap a neutral habit for another neutral habit. They are being asked to abandon the very strategy that made them successful. That is not a behavior change. That is asking someone to bet against the thing that has never let them down.

The instrument is wrong for the kind of change

Now watch what organizations actually do. Faced with the second kind of change — a change in mindset or behavior — they reach for the playbook built for the first. They announce the new state. They mandate it. They build the case, run the training, roll out the communications, and assign someone to manage the “change.” It’s the payroll-migration toolkit, scaled up and pointed at a transformation.

And it backfires, for two reasons that are worth separating because the field almost always collapses them into one word — resistance — and then mismanages both.

The first reason is about choice — who decided. People have a reliable response to having a choice taken from them: they move to take it back. And the counterintuitive part, which researchers have documented for decades, is that people will push back against an imposed option even when it’s the better one — because the pushback isn’t really about the option. It’s about who gets to author the decision. So the moment you announce that people will now think like innovators, or must embrace the new tools, you have converted something they might have chosen into something being done to them. You’ve manufactured the very pushback you’ll spend the next year managing.

The second reason is reinforcement — what pays off. People keep doing the old thing because, in the current environment, the old thing still works. The quarterly numbers still reward the leader’s command-and-control reflex. The expert’s hard-won judgment is still what gets them respect in the room. Routing supplies through the familiar back channel still gets the supplies. The behavior persists not from stubbornness but because it remains the rational move given the conditions. Communication won’t touch this. Training won’t touch this. A leadership message certainly won’t. As long as the environment still rewards the old behavior, the old behavior is the intelligent choice — and changing makes no sense.

These two pulls are different, which is exactly why one playbook can’t address both. Choice shows up even when the new way would genuinely serve the person, because it was imposed on them. Reinforcement shows up even when nothing was imposed at all — because the payoff never moved. Mandates make the first one worse. Pep talks ignore the second entirely. And the harder an organization pushes on either, the further underground the real behavior goes.

What this looks like right now

You don’t have to imagine the example, because everyone is living inside it. The most consequential adaptive change of the moment is the arrival of AI into knowledge work, and it is revealing many of the challenges this perspective would predict

The resistance here is not that people can’t operate the tools — the tools are, if anything, easy, and people are comfortable updating tools. A large global survey of full-time employees this past winter found that adoption itself is no longer the real challenge: people are broadly optimistic about AI, and many already use it every day, which tells you something important. If the holdout were about ability, training would fix it. It isn’t, so it won’t. What remains, once competence is off the table, is about identity. You are asking someone who spent decades building specialized judgment to work beside a system that performs a version of that judgment in seconds. The quiet question underneath isn’t what button do I click. It’s who am I now, if the thing I was prized for can be done without me? That is a threat to the self, and a person who feels that threat does not do their best, open, exploratory thinking. They protect themselves — they look for the exit, or the workaround.

And the organizational response has mostly been the old playbook: change-manage adoption through a comms plan. The predictable result is already visible. A Fortune report this spring described white-collar workers quietly refusing adoption mandates, with a trust gap so wide that executives and their employees seem to be describing different companies — leaders certain the tools are adequate and embraced, workers experiencing something else entirely. Meanwhile, the same people are using AI fluently in their own lives, in their own way, just not in the sanctioned form they were handed. People are adapting — only not to what they were told to adapt to. And that is the quiet signature of the whole error: the organization sees the surface compliance it asked for — the login, the attended training, the nod in the meeting — and mistakes it for the change it never actually got. Compliance produces the form of the new behavior. The substance never arrives, because substance was never something you could mandate.

Consider for a moment the ones who say nothing and don’t change. It’s easy to read their silence as a failure to get on board. But in the face of an identity threat, that silence is not resistance — it’s often the most useful signal you’ll get: a person telling you, in the only way available to them, that the new way hasn’t yet been made to make more sense than the old one. Miss that signal, and you lose twice. You lose what doesn’t show up on a balance sheet — leadership credibility, and the speed an organization only gets when its people adapt willingly. And you lose the person: someone who could have been brought along, written off for behaving exactly the way people behave, when a little more care in how the change was offered might have kept them.

The problem isn’t leadership, or the employees, or the change itself — it’s the approach. Any change that depends on a shift in mindset or behavior has to begin with a different question: not how do we make people adopt it, but are they ready, and would they choose it? You don’t tell a skeptical expert to believe in AI. You create conditions where they can use it to solve a problem they already care about, on their own terms, and you watch what happens. Nobody announces a transformation. It simply becomes the new normal, the way the smartphone did. You let them experience a different future before asking them to believe in it — which is, I think, as close as I can get to the heart of the thing.

I’ve spent years calling this covert disruption, and the name has stuck because it captures what’s actually happening: the change arrives sideways, through experience rather than instruction, and takes hold before anyone braces against it. Anyone who’s worked with me has heard the term. But the word carries an obvious charge, and it’s worth meeting head-on.

Why can’t the argument do it

If the block were a lack of information, you could argue that someone across it. But arguing harder is only more of the same imposition, and the harder you push a person whose sense of themselves is on the line, the more they brace. Under that kind of threat, people don’t weigh the better case; they defend the self the case seems aimed at, holding the inconvenient evidence to a standard nothing could meet. You’ve watched a room full of smart people do exactly this.

There’s a quieter reason too, and it’s worth only a sentence here because it opens onto something much larger than this essay: a great deal of what would change a person’s mind never reaches them at all. The biggest limitation in how people think isn’t bad thinking. It’s invisible thinking — the information their own filter set aside before they could consider it. You can’t argue someone past a filter they don’t know they’re using. You can only create an experience it can’t screen out.

Which is why the thing that works looks so indirect. A good coach rarely supplies the answer; they ask the question that lets the person arrive at it, because an insight you reach yourself arrives with your authorship intact, and what comes as your own discovery never trips the defenses an imposed conclusion would. I saw something new changes a person. Someone made me accept something hardens them.

Does covert equal manipulation?

So let’s not pretend the word doesn’t sting. If you’re shaping the conditions so that people change without quite noticing, isn’t covert just a softer word for manipulation, worse, manipulation with neuroscience behind it?

I think the discomfort dissolves once you see what “covert” actually refers to. It isn’t the intervention that’s hidden from the person. It’s the adaptation itself that is covert — the same way your growing ease with a new city was covert, the same way no one can name the moment a team became genuinely collaborative or a person became a leader. The change happens below awareness through repeated experience, and that’s simply how adaptation works, whether anyone designed for it or not. Nothing is being concealed from the person.

Manipulation overrides someone’s judgment to move them somewhere you’ve decided they should go. What I’m describing does the opposite: it removes the manufactured pressure — the mandate, the threat, the engineered buy-in — that was preventing the person from seeing clearly and choosing freely. It returns the choice to them. And it carries its own honesty test built in: if the new way won’t take hold no matter how well you’ve set the conditions, you stop, because the refusal is information. It’s telling you the change doesn’t actually serve the person — and pushing through that refusal is the precise moment you’d cross from creating conditions into coercing an outcome. The manipulator overrides the no. This approach treats the no as the truth it is.

What people are actually adapting to

So the instruction was never “manage the change.” It was never even “change the people.” People are already adapting — expertly, continuously, every day of their lives. They are adapting, right now, to the environment they’re actually in: the incentives that still reward the old behavior, the status that still attaches to the old expertise, the threat that still hangs over the new way. They are behaving intelligently in response to their conditions. They always are.

Which means the real work is not on the people at all. It’s on what they’re adapting to. Organizations pour enormous effort into engineering behaviors directly, when behavior is almost always the byproduct of the conditions people are standing in. Change the conditions — make the new behavior the one that genuinely solves a real problem, remove the imposition that turns a possibility into a threat — and the adaptation takes care of itself, in the only way adaptation ever really happens: quietly, on the person’s own terms, until one day it’s simply how things are done and no one can point to when it changed.

This is no longer a fringe view, which is worth saying for anyone who suspects it’s wishful. Even the largest workforce-research institutions have begun, in their most recent work, to say that the old model of managing change through one-off programs and top-down communication is no longer keeping pace — and that the organizations pulling ahead are the ones embedding adaptation into the daily flow of work and treating culture itself as the infrastructure of change, rather than something to announce. The establishment that sold the old playbook is quietly conceding it doesn’t work. The interesting question is no longer whether to change the conditions instead of the people. It’s whether we understand how — and build the capability to do it.

Organizations are where this gets ignored most expensively, which is why I’ve used them here. But the same quiet logic runs underneath a great deal else — the coach’s question instead of the coach’s advice, the teacher who lets a student reach the answer, the slow private way each of us revises who we thought we were. The organizational version is one application of something larger. It may be the most costly place we get it wrong, but it is not the only one, and it is not the deepest.

For now, it’s enough to notice the thing hiding in plain sight: the most powerful change is the kind nobody experiences as change at all. We already know this. We’ve each lived it a hundred times. We’ve just never thought to run an organization as though it were true.