The diagnostic
FrançaisHow is your AI transformation working?
Most AI readiness assessments return a score. This one returns a reading — three separate measures of where your organisation is off course, and which deviation is widest.
It takes about five minutes. Nothing is asked of you before you see the result: no email, no form, no call.
One instruction that matters: answer from what you observe, not from what you hope is true. An instrument calibrated to optimism reads zero deviation on a course that is badly wrong.
Process Reality
Are we doing the right things, in the right order, at a speed our organisation can actually sustain?
Human Transition
Are our people and leaders actually equipped — psychologically and structurally — to operate in a world that keeps shifting under their feet?
Governance Steering
When should we say no — or not yet — to an AI opportunity, even if it is technically exciting?
Three readings, not a score.
A composite number would let you feel broadly fine. These are measured separately because they fail separately.
- Process Reality
- Human Transition
- Governance Steering
Two axes read equally wide. Process Reality is shown below — the other is worth the same attention.
Two axes read equally wide. Human Transition is shown below — the other is worth the same attention.
Two axes read equally wide. Governance Steering is shown below — the other is worth the same attention.
Your widest deviation is on process reality.
The common pattern on this axis is an organisation whose documented processes and actual processes have drifted apart, where AI is being layered onto existing ways of working rather than prompting the question of whether those ways should exist at all. The tell is usually vocabulary: initiatives described in terms of speed and cost rather than in terms of what becomes possible.
This is not a technology problem, and more tooling will not close it. What closes it is mapping how value actually moves through the organisation — process by process, judgement point by judgement point — starting from what people actually do rather than from the org chart. Prioritisation comes after that mapping, not instead of it.
Your widest deviation is on the human transition.
The useful move here is to stop calling it resistance. Resistance is a label, not a diagnosis. What matters is which specific transition is failing: people may understand the change and not want it, want it and feel unable to act, or be holding on to something — expertise, standing, a way of working — that nobody has acknowledged they are being asked to give up. Each needs a different response, and generic leadership development addresses none of them.
The distinction is diagnosable. It requires hearing from leaders and from their teams separately, because the gap between those two accounts is usually where the answer sits.
Your widest deviation is on governance.
The pattern here is governance operating as a constraint function — reviewing, approving, declining — with no corresponding steering function. Such a board can usually say what it has rejected, but not what it has originated. AI reaches the agenda through external triggers: a competitor's announcement, a regulatory change, something in the press.
This is rarely a competence problem and almost never solved by another briefing. It is a question of whether the board has the standing and the confidence to ask generative questions — what should we be exploring that we are not? — alongside protective ones, and whether innovation and risk are permitted in the same conversation.
All three axes read on course.
That is a real result, and it is also the reading most worth testing. An instrument answered by one person returns one person's view, and a uniformly healthy view from a single position is the pattern most likely to shift when a second position is added.
This is not scepticism about your organisation. It is how the instrument works: you were asked to answer from what you observe rather than what you hope, and no self-assessment can check whether that held. A second reading, from someone at a different level, can — and if theirs matches yours, you have something considerably more solid than a single result.
All three axes read equally wide.
No single deviation dominates. That can mean the difficulty is genuinely even across process, people and governance — or that the instrument has been answered at a constant distance from the detail, which happens when one position cannot see all three clearly.
Either way the next step is the same. A second reading from a different level will separate them: evenness seen from one vantage point rarely survives contact with a second.
One position
This is one reading, from one position.
You answered as one person, from one vantage point. The most valuable signal in any diagnosis is rarely what one person reports — it is the distance between what different levels report. Executives and process owners. Leaders and their teams. The board and management.
A board that rates its own steering at four, where management rates it at two, has found something. Neither number is wrong. The gap between them is the finding.
Send this to two colleagues at different levels and compare your readings.
The starting point
A thirty-minute diagnostic conversation. No charge and no slides.
Your widest deviation reads on process reality. If that matches something you already suspected — or contradicts it — that is worth half an hour.
Your widest deviation reads on human transition. If that matches something you already suspected — or contradicts it — that is worth half an hour.
Your widest deviation reads on governance steering. If that matches something you already suspected — or contradicts it — that is worth half an hour.
Your reading is clean across all three. Whether it stays clean when someone else answers is the more interesting question, and that is worth half an hour.
All three axes read equally wide. Which one actually matters most is precisely what half an hour would establish.
Keep this reading
A copy of your three readings, sent once. Nothing else.