The Terraformer Manifesto
The difference is not whether we change. The difference is who holds the pen.
The fitting room
Somewhere right now, someone is being fitted for the future.
They are sitting in a training session about a tool they did not choose, learning a workflow designed by people who have never seen their work, being measured against a definition of “AI-ready” written in a vendor’s product roadmap. They are being told, kindly and repeatedly, to adapt. Reskill. Stay relevant. The world has changed; adjust yourself to it.
I refuse the premise.
Not the technology — I use it every day. I refuse the direction. Much of AI transformation points one way: take the human, reshape the human, until the human fits the machine-shaped world someone else deployed. The person is the variable. The environment is the constant.
A terraformer holds it the other way around.
The atmosphere, not the furniture
For much of my career, my profession — transformation, programmes, portfolios — has changed organisations by changing structures, processes, governance and systems. Serious work. But much of it shared one silent assumption: that the deeper conditions of work — what counts as expertise, what a deliverable is, what a team is for, what deserves a week of thought — were relatively fixed.
We moved the furniture. The air in the building stayed the same.
AI changes the air. It changes what knowing means, what making costs, what one person can carry, what can be delegated, what can be remembered, and how quickly an idea can become something that works. It reaches into the conditions of work themselves.
An atmosphere no one designs still gets designed — usually according to somebody else’s priorities.
The terraformer’s object of work is the environment itself. Not simply the organisation within its rules, but the rules. The aim is not adoption of the new world. It is authorship of it.
AI is not a new world humans must adapt to. It is material with which humans can build a new world.
Terraforming is the right word. You do not train settlers to breathe poison. You build an atmosphere worth breathing.
The inversion
The human is the fixed point. Everything else is the variable.
The point is not that humans should stop changing. We should evolve — we always have. A terraformer learns the new tools, tests them, fails with them and rebuilds. But there is a difference between evolving as an author — learning in order to shape what comes next — and being changed as material, reshaped to fit an environment somebody else has already decided to build.
The difference is not whether we change. The difference is who holds the pen.
The human — with their needs, dignity, accumulated craft and judgement — remains the point of reference. The AI layer gets shaped around them. Not because work should remain as it is, but because technology should be the material of transformation rather than its unquestioned destination.
The pen is not innocent. Humans disagree about pace, autonomy, risk and what dignity at work means. There is no single human around whom an environment can simply be optimised. Terraforming therefore includes negotiation: who gains control, what becomes invisible, what people are still expected to know, and what they remain able to challenge.
There is another difficulty: authorship is not evenly distributed. Many people are handed systems they did not choose and expected to adapt to them. If terraforming means anything, it cannot simply give more authorship to those who already have it. It has to create more room for people to shape the environments they work inside.
We all adapt where we must and author where we can. The ambition is to expand the second.
The maker returns
Years ago, I was transformation director at a bank. It was the kind of programme with a war room, a vendor army and status packs thick enough to stop a door. My job was the design: the new operating model. Other people owned the build.
The bank had become a corridor of handoffs — cases bouncing between teams, re-explained three times. My model meant one journey, one owner, decisions in hours rather than weeks.
And we delivered it. RACI signed, forums in the calendar, system live, training complete, workstream after workstream green for months. The closure report used the sacred phrase: “delivered to scope.”
Procedure successful. Patient dead.
What took me years to name was that nobody had obviously failed. Everybody had succeeded against their own scorecard. The PMO hit its milestones, the vendor passed acceptance, managers demonstrated compliance. Nobody’s scorecard asked the question the programme existed to answer: does a real case, today, actually travel the new way?
I learned it in one week. Monday, steering committee: a wall of green heatmaps. Thursday, on the floor, I asked a team lead to walk me through one live case. The story frayed at every step. Then he looked at the wall behind him — my operating model, printed, a little faded at the corners — and said: “On the poster, yes. In real life, we still know who to email to make it move.”
The artefact had shipped. The atmosphere had not.
And I could not fix that with another deck. I did not lack skill. I lacked hands. Seniority had made me the author of a drawing other people performed — each faithful to their part, none responsible for the whole living thing.
That distance between the person designing change and the hands constructing it is one of the structural weaknesses of my profession.
AI has begun to narrow that distance. I can now design parts of a new way of working and construct parts of its machinery myself. Not slideware about agents — agents. Built in my own work first. Used. Broken. Changed. Used again.
That is what The Lab is for: not to prove that I know where all of this is going, but to make the claims testable. Prediction before evidence. Verdict after. Failures included.
The philosophy can be argued with. The practice can be watched.
AI is unusual among technologies in being genuinely plastic: material that can be worked, not weather that must be endured. A practitioner who can build can shape that layer around this team, this practice, these people: the weekly rhythm, what counts as finished, where verification lives, what gets remembered and by whom, what deserves human attention, what no longer deserves to exist.
Human-centred AI without the ability to shape the environment is a wish. With that ability, it starts becoming a discipline.
And the discipline should begin close to home. Before prescribing an AI environment to someone else, seriously inhabit one yourself. Otherwise we risk becoming travel agents for a planet we have never visited.
The terraformer is not a new species. It is what a transformation practitioner starts to become when the maker returns.
The weight
One more thing.
You can undo a reorganisation. You cannot entirely un-terraform.
The tools may be reversible. An agent can be switched off. A check can be reinstated. But what people quietly stopped doing does not necessarily return on command: the verification muscle, the craft memory, the habit of knowing rather than retrieving, the confidence to challenge the answer.
You can uninstall the software. You cannot simply uninstall the atrophy.
And the downstream is wider than the org chart. An atmosphere runs on energy, infrastructure and other people’s resources. It creates dependencies its inhabitants may never have chosen. One environment cannot be called improved if it was built by quietly degrading another, or if the people living inside it can no longer understand it, question it or change it.
Whoever shapes an environment other humans will live inside carries responsibility for more than whether the system works. They carry responsibility for what it strengthens, what it weakens, what it allows to disappear and, eventually, for what it makes people become.
If you think an organisation can simply walk back from a changed atmosphere, argue with me.
That is what this margin is for.
Terraformisti
I wrote this because I suspect people are already doing this work without a name for it.
Transformation directors who have started building again. Programme and portfolio practitioners shaping the AI layer around their people instead of the reverse. People learning the new physics in their own work before prescribing it to others. People who see AI not as a destination to which humans must migrate, but as material from which a different environment can be built.
In my language there is a word forming for them. I like how it sounds in the plural: terraformisti.
Not a club. Not a title. A stance.
AI is being presented as a world we must learn to inhabit.
I think we should ask a different question: what kind of world do we want to build now that we have this material?
The atmosphere is being written right now, with us or without us.
I would rather hold the pen.