The (AI) Change Starts With You

“There is a crack in everything, that’s how the light gets in.” — Leonard Cohen, Anthem

• • •

Try to imagine that it is the year 1869.

You are a young researcher. You have just returned from Paris, where you studied Pasteur’s work on fermentation and infection. You have seen Lister’s results in Glasgow — how the simple use of carbolic acid as an antiseptic reduced surgical mortality from 45% to 15%. You have the evidence. You understand the mechanism. And now you have been given an assignment.

You must stand before a room of the most distinguished surgeons in the country and explain to them that because they do not wash their hands or sterilise their instruments, they sew death into every wound. You must persuade them to abandon most of what they have been taught, to question the wisdom they have accumulated over their distinguished careers, and to rebuild their understanding of the practice of medicine around a new theory. The theory of germs.

Do you think you could do it? Do you think you could convince them? Do you believe they will be glad to hear you?

Myron Tribus posed this scenario in 1992. He was not writing about medicine. He was writing about management.

The parallel is uncomfortable and precise. The surgeons were not bad people. They were sincere, experienced, and respected. They had saved lives. They had earned their authority. And with every operation they performed with unwashed hands, they killed patients — not because they intended harm, but because the prevailing theory of their profession prevented them from seeing the mechanism that connected their behaviour to the outcome. They had a theory. They called it “mal-aria” — bad air. And that theory caused them to look in the wrong places for wrong answers to their most pressing problems.

Tribus’s point was this: management is in the same position today that medicine was before the germ theory. Not everything managers do is wrong — just most things. And the managers, like the surgeons, are not to blame. They are prisoners of their upbringing, their culture, and the state of knowledge of their profession.

If you have read the first five articles in this series, you have already met the modern version of this problem. You have seen how AI applied without understanding does the wrong things faster. How the system you build amplifies whatever thinking you bring to it. How the truth about your organisation lives in the work, not in the strategy. How people pay the price when leaders optimise for the wrong thing. And how the hardest barrier to change is not the system around you — it is the assumptions you carry about yourself.

This final article is about the question that remains: now that you understand, what do you do? And more precisely — how do you help others understand, when the very act of telling them is the thing most likely to prevent it?

• • •

What happens when purpose shifts

Before we get to that question, I want to widen the lens. Because the pattern we have been tracing through this series — well-intentioned people causing harm because they do not understand the system — is not limited to your organisation. It is playing out at the largest scale imaginable.

Consider social media. Facebook did not set out to damage teenagers. The founders wanted to connect people. But somewhere along the way, the purpose shifted. The user was no longer the customer — the advertiser was. The metric that mattered was no longer “are we helping people connect?” but “how long can we keep them on the platform?” The system optimised for engagement, which meant optimising for outrage, anxiety, and compulsive checking. Facebook’s own researchers found that Instagram made body image issues worse for one in three teen girls. When the company introduced safety measures during elections, they dissolved them afterwards — because the measures undermined virality and growth.

The damage was not a bug. It was the system working perfectly against a purpose nobody had the honesty to examine.

Or consider Boeing. For decades, Boeing was an engineering company. The culture was: build the best aircraft, and the money follows. Then came the 1997 merger with McDonnell Douglas. The incoming leadership was explicit about the intent: run it like a business, not an engineering firm. Headquarters moved from Seattle to Chicago — physically separating the executives from the factory floor. Between 1998 and 2018, Boeing spent $61 billion on stock buybacks — 82% of its profits. The metric shifted from engineering excellence to shareholder return and speed to market.

When Boeing designed the 737 MAX to compete with the Airbus A320neo, the governing constraint was not safety — it was cost. The aircraft needed to keep the same pilot type rating as the previous 737 to avoid expensive retraining for airlines. To make this work, they added a software system called MCAS to mask how differently the new aircraft flew. Engineers and test pilots raised concerns. One pilot reported that the system was “running rampant” in the simulator. Internal communications described the aircraft as “designed by clowns, supervised by monkeys.” The concerns were overruled by the business logic. The FAA had delegated safety certification to Boeing itself.

346 people died. Not because anyone at Boeing intended harm. Because the system was optimising for exactly what it had been designed to optimise for — and the people inside it were rational actors responding to the metrics they were measured by.*

The argument chain is always the same. Shift purpose away from the people you serve. Optimise for the wrong thing. The system produces exactly what it was designed to produce. And the consequences are devastating — first for the people who depend on you, and ultimately for the organisation itself, because a system that destroys trust cannot sustain itself.

This is Tribus’s germ theory, playing out in our time. Well-intentioned people, operating within the best knowledge of their profession, causing harm they cannot see — because their theory prevents them from looking in the right place.

• • •

But this article is not about Facebook or Boeing

It is about you. The leader who has read five articles, who has perhaps started to see the pattern in your own organisation, and who is now asking the hardest question: what do I do with what I see?

The last article ended with a paradox: the clearer you see the potential for change, the more tempting it is to do the one thing that prevents others from seeing it too — to tell them. To present the evidence, explain the argument, and ask them to get on board. After all, you have done the work. You understand the system. Surely the efficient thing is to share the conclusion and move forward.

This is the most seductive mistake a leader can make. And it is the same mistake the young researcher in 1869 is about to make — standing before the surgeons, armed with evidence, ready to explain. The surgeons will not be glad to hear it. Not because they are stupid. Not because they do not care. But because what the researcher is really asking them to do is abandon the framework through which they understand their entire professional identity. No amount of evidence can make that feel safe.

• • •

Humility as prerequisite

Chris Argyris spent decades studying why intelligent, successful people are often the worst at learning. His answer was not about intelligence — it was about ego. Or more precisely: it was about the defences that ego constructs, invisibly, to protect itself from the discomfort of being wrong.

Argyris called these defences “defensive routines” — patterns of thinking and behaving that shield us from embarrassment or threat while simultaneously preventing us from learning. What makes them so powerful is that they operate below awareness. The more successful you have been, the more practised your defences, and the less likely you are to notice them at work.

In the previous article, I introduced Argyris’s observation that most leaders espouse one set of values — openness, collaboration, learning, inquiry — while their actual behaviour reveals another: maintain control, win the argument, suppress negative feelings, appear rational. He called this the gap between “espoused theory” and “theory-in-use” — or, put plainly: the gap between what you say and what you do. It is not hypocrisy. It is the natural consequence of spending a career in a system that rewarded control and certainty and now needs you to be curious and uncertain.

But there is a deeper point that Article 5 only touched on, and it is the one that matters most for this conversation.

Argyris identified three conditions that must be present for genuine organisational learning to occur. Three conditions — not twelve principles, not a framework with quadrants, not a maturity model. Three.

The first is valid information. Everyone involved must have access to the real picture — not the version that has been filtered, curated, or softened to protect feelings or reputations. Valid information means you can see what is actually happening in the work: the real demand, the real flow, the real outcomes measured against the real purpose. It means the data has not been tortured into telling a convenient story.

The second is free and informed choice. People participate because they understand the situation and choose to engage — not because they have been told to, not because it is on their objectives, not because resistance has been made career-limiting. Informed choice means people have enough understanding of the evidence and enough freedom to say “I disagree” or “I need to see more” without consequence. The moment compliance replaces choice, learning stops.

The third is internal commitment. People believe in the direction because they arrived at it through their own understanding — not because it was mandated, not because a compelling presentation made it feel inevitable, not because the CEO said this is where we are going. Internal commitment cannot be designed. It cannot be accelerated. It emerges — or it does not — depending on whether the first two conditions are genuinely present.

And sometimes free and informed choice means helping someone arrive at a different conclusion: this is not for me. That is a legitimate outcome, and it should be met with grace. What cannot be sustained is staying and refusing to act on what the work is teaching you — what your role requires you to learn. That is not choice. That is avoidance.

These sound reasonable. Most leaders would say they agree with them. And this is precisely the trap.

Do you share valid information — or do you filter it? Do you create the conditions for free and informed choice — or do you present a conclusion and ask people to get on board? Do you allow commitment to emerge from understanding — or do you design programmes that produce compliance dressed up as buy-in?

If you are honest — and the whole point of this series is that honesty is the prerequisite — you will recognise moments where your behaviour has not matched your values. Not because you were acting in bad faith. Because humility is the hardest thing the system never taught you. The system taught you to know, to decide, to lead from the front. What is required now is the capacity to not know, to inquire, and to create the space for others to arrive at their own understanding. That is not weakness. It is a fundamentally different kind of strength. And it requires you to let go of the thing your ego most wants to hold onto — the feeling of being the one who sees it clearly.

• • •

The questions that change everything

So what does this look like in practice? If you cannot tell or explain people into understanding, what can you do?

You can ask different questions.

Not different in the sense of cleverer or more provocative. Different in the sense that they come from a fundamentally different set of assumptions about how organisations work. Most of the questions leaders ask are single-loop questions — they accept the current framework and try to optimise within it. Are we hitting our targets? Are our people performing? Is the process efficient? These questions keep you inside the existing logic. They cannot reveal whether the logic itself is the problem.

The questions I am about to describe work differently. They do not ask you to optimise. They ask you to understand. And in doing so, they surface the thinking that designed your current system — which is the first step toward changing it.

Start here: what is the purpose of this system — from the perspective of the people it serves? Not your mission statement. Not what the strategy document says. What would your customers, citizens, or service users say the purpose should be, based on what they actually need from you?

Then: what is the nature of demand? Why do people contact you? What do they want? And critically — how much of that demand is caused by your own failure to do something, or to do something right, the first time? When you answer this question honestly, you find out what your system is actually producing — as distinct from what you think it is producing.

Then: how does the work flow? Not the process map on the wall — the actual journey a piece of work takes from the moment it arrives to the moment it is resolved. Every handoff, every queue, every rework loop, every moment of waiting. Follow it. Measure the time. Count the touches. You will see waste you did not know existed, not because people are lazy, but because the design of the system makes waste inevitable. You will see the work cross functions and you follow it — it does not care about your structure. That is the point.

Then: what measures would tell you how well you are serving purpose? Not activity. Not output. Not compliance. Purpose. If the purpose of a claims process is to return customers to their pre-loss position, then measure the time and experience of that return. If the purpose of your health system is to help people live the life they want — do you actually do that? Do you see the whole person, or only the diagnosis? When you measure against purpose, you see things that activity metrics hide — and you stop optimising for things that do not matter.

And finally: what are the system conditions that are producing the performance you see? What targets, incentives, procedures, IT constraints, management practices, and structural arrangements are shaping how people do the work? This is the question that takes you to the governing variables — the assumptions built into the design of the system itself. Not who is performing well or badly, but what is determining performance.

Notice what these questions do. They do not ask you to improve the existing system. They ask you to understand it — on its own terms, from the outside in. And in doing so, they reveal the thinking that built the system in the first place. That is double-loop learning: not “are we doing things right?” but “are we doing the right things — and what thinking led us to believe these were the right things?”

Here is where it connects to everything this series has argued. When you ask “why are we using targets?” — really ask it, and follow the logic — you arrive at the underlying assumption: targets exist to control the work and the worker. That is the theory-in-use, regardless of what the espoused theory says about empowerment and trust. And once you see it, a different world opens up. What if the purpose of measurement is not control but learning? What if you measured to understand rather than to manage? What if the question was not “are they hitting their numbers?” but “what are we learning about how this system serves the people who depend on it?”

When you make that shift — and it is a shift in thinking, not a technique — something remarkable happens. The people closest to the work start to see things they could not see before. And because they can see the system, they can improve it. Not because you told them to. Because the evidence is in front of them and the conditions allow them to act on it. The conversation between leader and worker changes. Instead of “why didn’t you hit your target?” it becomes “what do you need me to change so you can do a better job?” The leader’s role is no longer to control performance. It is to remove the obstacles that prevent good work.

And here is the quiet surprise: when this works — when the system has genuine autonomy and purpose — you may find you are not that busy. Not because you have abdicated. Because the system works, and the people in it know what to do.*

These questions were not invented by me. They were developed by John Seddon, and the method built around them — the Vanguard Method — has been tested across hundreds of organisations in the public and private sector, in multiple countries and languages. I have used it for over a decade. The reason it works is precisely what this article has been arguing: it does not tell leaders what is wrong. It puts them in front of the evidence and lets them see it for themselves. It creates the conditions for valid information, free and informed choice, and internal commitment — Argyris’s three prerequisites — not as philosophy, but as practice.

The method works because it embodies the very thing it teaches.

• • •

One more thing

There is a consequence of building a system this way that is easy to overlook, and I want to name it before we close.

When people can see the purpose of their work, when they have the information to understand how the system performs, when they have the freedom to act on what they see and the support to remove obstacles — something happens that no engagement survey or motivation programme has ever reliably produced. People become motivated. Not because you motivated them. Because the conditions for motivation were built into the system.

The question “how do we motivate our people?” disappears. It was never the right question. The right question was always: what conditions are we creating — and do those conditions enable people to do meaningful work?

• • •

Sorry about that

If you have read this far — all six articles — you may have noticed something.

This was never really about AI.

Sorry about that. Well — not really sorry.

AI is in the title of every article. It is what drew you in. It is urgent, it is everywhere, and it is the reason most leaders are paying attention to their organisations right now. But the journey this series has taken you on — from understanding the problem, to seeing your system, to recognising the human cost, to facing your own assumptions, to asking fundamentally different questions about how you lead — none of that required AI. All of it is about something older and deeper: whether you have the courage to ground your decisions in what is true, and the method to find out what that is.

The title of this article asks: how do we make the change? The “(AI)” is in brackets — and now you know why. Because the change that matters is not the technology. The change that matters is in the thinking.

And here is the thing about AI: once you have done this work — once you understand your demand, your flow, your purpose, your system conditions — AI becomes genuinely extraordinary. Not as a replacement for understanding, but as an amplifier of it. The same tool, the same power, an entirely different outcome. Because you understood what you were pointing it at before you picked it up.

• • •

The surgeons learned

Tribus’s story does not end in 1869. It ends with this: the germ theory won. Not quickly. Not without resistance. Not because someone gave a better lecture. It won because the evidence became undeniable to those who were willing to look. Lister’s patients survived. The surgeons who washed their hands stopped killing people. And eventually — painfully, generationally — the profession changed.

Not everything the old doctors did was wrong. Just most things. And what they were taught was not good enough. Some things they did were downright dangerous and harmful. They learned. So can you.

The evidence is in front of you. The method exists. The questions are waiting. The only thing that remains is whether you are willing to look — and to let others look with you.

---

If this series has changed how you think about your organisation, your leadership, or the role of AI — I would love to hear from you.

You can reach me at jonas@vanguard-consult.dk or connect with me on LinkedIn.

---

SOURCES AND ACKNOWLEDGEMENTS

• Myron Tribus (1921–2016) was a physicist, engineer, and one of the leading interpreters of W. Edwards Deming’s philosophy. The Germ Theory of Management was published by SPC Press in 1992. All Tribus quotes in this article are verified against the original text.
• Boeing 737 MAX: the factual account draws on the US House Committee on Transportation and Infrastructure report (2020), PBS Frontline reporting, and the Department of Transportation Inspector General’s findings on FAA certification. Boeing pleaded guilty to criminal fraud conspiracy in July 2024. *For a vivid account, see the documentary Downfall: The Case Against Boeing (dir. Rory Kennedy, 2022), available on Netflix.
• The social media and mental health evidence draws on the US Surgeon General’s Advisory on Youth Mental Health (May 2023) and Frances Haugen’s disclosures of internal Facebook research (2021).
• Chris Argyris’s work on defensive routines, espoused theory vs. theory-in-use, and the three conditions for organisational learning draws primarily from Organizational Learning II (with Donald Schön, 1996) and On Organizational Learning (1999).
• Gerard Egan’s concept of the “alternative world” — helping people envision what could be different — is from The Skilled Helper (multiple editions, first published 1975).
• *The observation that leaders in well-functioning systems may find themselves with surprisingly little to do echoes Ricardo Semler’s account in Maverick (1993) and The Seven-Day Weekend (2003).
• John Seddon developed the Vanguard Method and the systems perspective on service organisations. The method questions described in this article are central to his work. See Freedom from Command and Control (2003) and The Whitehall Effect (2014).
• The series concept of failure demand — demand caused by a failure to do something or do something right for the customer — was originated by John Seddon.

---

Part 6 of a 6-part series on AI and the systems perspective.

Part 1: AI Won’t Save Your Organisation
Part 2: AI Does Not Care What It Amplifies
Part 3: AI Can Wait, Go and Look
Part 4: The Matchstick and the Body
Part 5: The Man Who Loved Dragons
Part 6: This article.

---

This article was written with AI assistance — which felt appropriate, given the subject matter.