Leadership in AI

This is the third of these letters — and the most important. The first described a world that has reset across four dimensions at once, and the second walked through the hardest call on the board’s desk — whether, and how, to commit to the AI journey. That letter ended on a line: not a technology — a journey that only leadership can carry.

This letter is about that leadership. Because a decision, however carefully made, transforms nothing by itself. Between the board’s resolve and the organisation’s reality stands a layer of people — the executives and senior leaders — who take that direction and translate it into action and make it real, through their teams, in conditions no one has seen before. Everything now depends on them. Much is written about this layer — usually as a bottleneck to be unblocked or a cohort to be upskilled. Very little is written for them — for the position they actually stand in and the weight they actually carry.

That is what this letter tries to address: the role you as a leader play in the AI transformation — and what it means for you as an individual, for your people, and for the impact you can create on the organisation. Some of what follows may already be your experience; for others, it may still be ahead — many organisations are in the early stages of their decisions on AI. Either way, terrain is easier to cross when someone has a map, and that is what this letter tries to be.

OneThe Role of the Leader

There is a particular moment when a transformation stops being a strategy and becomes a person’s Monday.

The board has decided. The townhall has happened; the words AI and future and journey have been said from the front of the room, and said well. And somewhere between that afternoon and the next operating review, the thing quietly changes hands. It stops belonging to the deck. It starts belonging to you.

From one direction comes the mandate. Deliver this. Move your function into a new way of working. Find the outcomes. The direction is set with conviction — and yet, if you observe honestly, it is a direction, not the destination. Where this journey ends — what the function looks like in three years, which roles it holds, what the work even is — is not fully written down anywhere, because no one can write it yet. The technology is moving too fast for anyone to finalise the end point.

You are being asked to lead people on a journey where you do not know the destination.

From the other direction come the questions. Not in the townhall — people rarely ask the real ones there. They come later, and smaller. In the corridor after a team meeting, in the last five minutes of a one-on-one, after the project talk is done and there is a pause. One of your best people — someone you hired, someone whose wedding you attended — looks at you and asks, in whatever words they choose: “Should I be worried?”

And you hear yourself beginning an answer you are not certain you know.

That is the position of the leader. The decision arrives from one side; the questions arrive from the other; and you stand where they meet, expected to hold both. Leaders have always stood there — that is what their job is. What has changed is that the two things a leader could always rely on in that spot — knowing roughly the destination, and knowing what to say to the people who ask — have both, at the same time, become genuinely hard to come by.

Have we seen this before?

It is worth pausing on a memory many of us carry, because a smaller version of this has happened inside our organisations before, and what it did — and did not do — explains a great deal about now.

If you came up through technology, you lived it. You remember when agile arrived — when the long project plans, the Gantt charts, and the weekly status packs gave way to small teams working in short cycles, and new words moved into the building. You also remember what it did to a particular group of people — the project managers. Capable, experienced professionals, whose value sat in the plan, the status, the control, watched the new way of working quietly stop needing the thing they were best at. No one declared their role over. It simply mattered less each quarter. Some found new footing as the roles reformed. Some moved elsewhere. It was not a small thing for the people it touched; for many of them it was the hardest passage of their working lives.

The engineers, notice, were touched far less. Their craft survived intact; they wrote code before and they wrote code after, albeit in shorter cycles. The shift fell hardest not on the people doing the work but on the people managing it — the layer whose value was coordination, oversight, the plan.

If you led elsewhere in the business, you may remember that period differently — mostly as vocabulary drifting out of the technology floor. Stand-ups, sprints, scrums. It changed how software reached you, and little else about your world. And that difference in memory is itself the important fact: you could afford to pay little attention to it because it was contained. It happened on one floor of the building while every other floor stood still. However hard it was for the people inside that floor, there was solid ground a few steps away — a stable organisation around them, senior leaders operating as before, a settled world to find new footing in.

Now hold that memory against AI. This time the change does not stop at one floor, one function, one method. It reaches how work is done in every part of the organisation at once — and beneath the ways of working, it reaches the operating premise of the business itself, which is what the first two of these letters were about. The layer it presses hardest is, again, the managing and the leading layer. But the thing that made the last shift survivable is missing: there is no stable rest-of-the-building this time. No settled floor a few steps away. The organisation that would normally give a leader their footing while they lead others through change — that organisation is itself what is changing.

Which brings us to the real question of this letter. Not how should leaders change — there is time for that. The first question is more basic:

When everything a leader steadies themselves against is in motion, what exactly are they standing on?

TwoWhat a Leader Stands On

Strip leadership to the core — beneath the titles and job descriptions — and it comes down to three aspects, built over a career, mostly without anyone naming them. I call them Credibility, Clarity, and Compassion — what people trust you for, what people look to you to provide, and what binds your people to you.

Credibility

Your credibility was earned the long way: by having seen and done it before. Thirty years — or fifteen, or twenty — of situations met and handled, patterns learned, judgement sharpened against real events. When you speak in a room, people listen because your experience has been right, again and again. That is not vanity; it is the honest basis on which organisations decide whom to follow.

Here is what has changed — because it is not that your experience has stopped mattering. It is that the situations now arriving have no earlier version. Not for you; for anyone.

The force producing them is new in kind — intelligence itself being manufactured, as the first of these letters put it — and so the instinct that has served you for decades keeps returning answers that may not fit the current context. The confidence you felt while answering drops — a little at first, and more as the results start showing on the ground.

You can feel the gap yourself, sometimes midway through answering: the pattern arrives, and some quiet part of you knows the pattern is from a different world.

And your people can feel it too — not because you have lost anything, but because they are living in the same newness. The team that once settled the moment you said I’ve seen this before, here’s how it goes now knows, as well as you do, that no one has seen this before. The room still respects you, but the sentence that used to close discussions has stopped closing them — through no fault of the person saying it.

Clarity

If credibility is what people trust you for, clarity is what they come to you to get. It has always been one of the leader’s quiet duties: taking the strategy from above and turning it into here is what this means for us — a direction and a destination that people can plan a year around, and underneath it, plan a life around. You have done this so many times it barely registers as work. Absorb the ambiguity at your level; pass down something steadier than what you have received.

Now look at what this moment asks. The direction handed to you is honest but unfinished — a journey whose destination moves as the technology moves and the organisation learns. This is no one’s failure of nerve; it is the truthful condition, all the way to the top. The board itself, as the last letter showed, is committing to a road it cannot fully map.

Which means what reaches you is provisional — and the thing you have always done, turning what you received into something steadier, has nothing settled to work with. Your people come to you for clarity that no one above you possesses either.

So you stand in the operating review, or at the front of your own townhall, doing something you never had to do: describing a destination while knowing it will change, committing to a plan while knowing it is written in pencil and will change. It can feel, from the inside, like a kind of pretending. It is not — it is what leading honestly looks like when the map is being drawn while you walk. But no one trained you for it.

Compassion

The third and most important trait of a leader — the one that, for many leaders, sits deepest. Underneath the delivery and the numbers, the part of the role you may privately value most: you look after your people. You know them. You know who is carrying a sick parent, who just took on a mortgage, who is quietly brilliant and two years from ready. Over the years, an unspoken bargain has formed: do good work, stay with me, and I’ll take care of you. You have kept that bargain often enough that your people believe it. That belief — more than any org chart — is what makes them yours.

This is where AI presses hardest. Because the question coming toward you is not about strategy. It is the one from the corridor: should I be worried? And for the first time in your leading life, both of the answers you have always had — where we are going, and how it will be in the future — are out of reach. You cannot say no, you will be fine — because you do not know that, and false assurance breaks the trust you have formed with them. Nor can you simply say you do not know — the hard truth — because that only deepens their anxiety. You can always say I don’t know about strategy, and it is fine. But to say I don’t know to a question about whether they have a place in the future is never easy.

Feel the precision of the trap here. The question sits exactly where honesty and comfort — the two things you have always been able to offer — can no longer be offered at the same time. Your knowledge of your people, usually your greatest asset, sharpens the difficulty: you can read exactly how afraid they are. The bargain you spent years keeping is being called in at the precise moment you have the least power to honour it — because the decisions that will determine who has a place may not be yours, and may not be anyone’s yet.

So, what is a leader going through? Credibility — asked to work without a precedent. Clarity — asked for when nothing above you is settled. Compassion — asked a question you cannot answer either way. Any one of these would make a difficult period. All three at once, with no still ground nearby — that is the actual weight of AI for the leading layer. Most writing about AI does not admit it — and leaders carry it silently, because they do not get corridors to ask their questions in.

I want to emphasise — none of this is a failing of the leader. A leader who feels all of these — the lack of answers, the pencil-written plans, the unanswerable questions — is not behind, and not lacking. They are standing exactly where the reset puts them: trained superbly for a world that held still while they led, now asked to lead in one that will not.

Which raises the question that the next section exists to answer. Not what should leaders do about all this — but the one that comes first: what should be done for them?

ThreeWhat Institutions Owe Their Leaders

If you sit on the board, or in the CXO chair, the previous sections were about your people. The transformation you have decided on will be carried well or badly depending on the condition of these leaders. If you make three commitments, you help resolve the three weights they are carrying — and the chances of the outcomes you have committed to rise with them.

1 · Give them a truth they can stand on. The hardest of the three weights is the question about people’s futures — and the deepest mistake an institution can make is to leave each leader alone to improvise an answer to it. The answer to should I be worried cannot be manufactured by a manager in a corridor; it can only be given, honestly, by the institution — in how it decides, early and from the top, to speak about the journey and what it means for its people. When the organisation says openly: the destination will evolve; here is what we know, here is what we do not, here is what we commit to about how we will treat people as we learn — every leader down the line is handed something they can honestly repeat. Their clarity problem eases, because what reaches them is no longer a certainty they must fake but an honesty they can extend. And the corridor question changes shape: the leader is no longer the sole author of an impossible answer, but the local voice of a truthful institutional one.

The shape of that honest answer was set out in the last letter: truth told as it becomes clear, including the parts that are hard to hear; reskilling that makes people genuinely more marketable in the AI world, whether or not their place here endures; and, if exits must come, a humane approach in which people share in the gains the change creates. What that answer should contain in full — the commitments an organisation can honestly make to its people — is the subject of a future letter, as I want to keep this one focused on the leaders. What matters here is who owns the question. It is not the leader in the corridor. It is you.

2 · Let them learn without spending their credibility. A leader whose standing was built on having seen and done it before faces a challenge in learning visibly: under the old rules, to learn in the open is to admit the experience no longer covers it. So many will try to learn privately, or to project a mastery that is not there yet — both of which slow everything. The institutions that are handling this well have broken this bind in the simplest way: their most senior people learn first, and in the open.

When a CXO is seen learning — asking basic questions in the open, being new at something — it becomes safe for every leader below to do the same. Only the top of the house can make that safe. And the need is not hypothetical: organisations everywhere are training people to use the technology, and almost no one is preparing leaders for what leading in AI asks.

And what does developing the leader actually mean? Not more content about the technology — briefings and licences are the easy part. It means building the two capacities everything else stands on. The first is the ability to think in whole systems (systems thinking) — to see how a decision ripples through people, technology, customers, capital, and risk before acting, because in a cascading world the second- and third-order consequences are usually larger than the first. The second is the strength to sit in not-knowing with composure (emotional intelligence) — to feel the weight of a decision without being ruled by it, and to hold others steady through a fear privately shared. Head and heart. Build those two, and the working capabilities this era asks for — reading a shifting landscape continuously, discerning what to preserve and what to rebuild, leading whole outcomes in flatter structures, carrying teams through uncertainty — have ground to stand on. Skip them, and every course is a coat of paint.

3 · Give them a hand in drawing the map. The weight on clarity is lightest when the leader is not merely receiving a provisional direction but helping to shape it. A destination that will evolve is a burden to announce and a different thing entirely to author. Bring the leading layer into the drawing of the journey — not consulted after the fact, but present where the direction is set, shaping the route for their part of the organisation, with real licence to adjust as the ground moves. A leader who helped draw the map can stand in front of their people and say here is where we are trying to go, and I had a hand in choosing the way — and mean it. The plan is still in pencil. But it is their pencil. That difference — between carrying someone else’s uncertainty and carrying your own shared, chosen risk — is the difference between a leader who transmits anxiety and one who transmits resolve.

Three commitments: a truth they can stand on, learning that does not cost them their standing, a hand in drawing the map. None of these require you — the CXO — to have certainty you do not have. All require something harder: going first.

And what do these three commitments quietly do? They redefine the three things this letter began with. Credibility stops meaning I have seen this before and becomes I can find the way through what none of us has seen. Clarity stops meaning I have the answers and becomes I can hold the direction steady while the answers form. And compassion stops meaning I can promise you it will be alright and becomes I will tell you the truth, and you will not face it alone. A leader remade in those terms has not been diminished by the loss of old certainties. They are closer to what leadership always was, underneath.

The permission to not-know is set by whether you — the CXO — model it; the safety of open learning is set by whether you are seen doing it; the honesty about the journey is set by whether you speak it before it is comfortable, and before you know all the answers.

Which is the same truth every one of these letters keeps arriving at, one level deeper each time: it is owned from the top, or it is not owned at all.

FourWhat You Owe Yourself

The institution owes you much of what this moment demands, and the previous section was addressed to it. This one is about what you can do for yourself. However much the reset changes, what you do with your own hands and your own time is still governed by you.

And here is the thing about AI: it cannot be understood by reading about it. Every previous shift could be studied from a distance — you could learn what the internet meant from people who understood it; the impact of digital on a business could be learnt and seen from a distance. AI reveals itself only in use. An hour spent working with it teaches more than a month of reports about it — and the reports cannot even stay current, given the rate of change. Which means the way back to solid footing does not run through a course catalogue or a transformation programme, however good they can be. It runs through your own desk.

So use it — and start with your own life, where nothing is at stake and no one is watching. Build with it the small system that helps manage your finances. Take the messiest part of your working week and rebuild it with the technology beside you. Prepare with it for a hard conversation. Take the operating review deck you are about to present and have it argue against you — you will meet the hard questions before your leader and peers ask them. Learn the subject you have been postponing for years. Sit with your children and watch how they think with it — those hours will teach you as much about the coming decade as any conference will. Make one small thing that solves one real problem you actually have. None of this needs permission, a programme, or a budget line. It needs you and your time.

And watch what it quietly returns to you. Credibility — because you become, in your own right, someone with present capability in the new world; not briefed on it, but capable in it. Clarity — because your view of what this technology is and is not stops being secondhand; you will have formed it yourself, from contact, and a view formed from contact holds steady in a way a borrowed one never does. And the corridor question — you still cannot promise what no one can promise. But you stand in it differently: as someone who has walked into the new country and found it navigable, and who can show the person in front of you how to take their own first step, because you took yours.

Leaders have never really led from certainty, whatever it looked like from the outside. They led from going first. That has not changed, and it does not require anyone’s permission. Independent of what your institution owes you, you owe it to yourself to take that first step. And I hope you do.