Enable
Adoption and change management
The part that decides whether any of it was worth building: the people who have to work differently, the managers who decide whether they will, and the evidence that your AI literacy obligation has actually been met.
Duration
8–16 weeks, usually alongside a first build
Built on
EU AI Act · ISO/IEC 42001
Indicative price
On request
Who this is for
COO
Has a working system and flat usage.
HR or transformation lead
Owns capability building and has no baseline to build from.
Head of a function
Team is polite about the new tool and has quietly gone back to the spreadsheet.
Compliance officer
Has to evidence AI literacy by role, not one module for everybody.
The model works. Nobody is using it.
The most expensive AI failure is not a system that does not work. It is a system that works, was paid for, and sits at eleven per cent usage while the work carries on as before. Nothing gets escalated, because nothing has broken — the old process is still there, and it is still allowed.
People do not resist AI in the abstract. They resist an unclear answer to three questions: what happens to my job, what happens when it gets something wrong and I acted on it, and who decided this. Left unanswered, those questions get answered informally, usually in the least generous version available. Answering them plainly and early costs less than any communications campaign afterwards.
Middle management is the hinge. Team leaders decide in practice whether a new way of working is enforced, tolerated or ignored, and they are typically handed the tool with no change to their targets. If a manager's numbers are unchanged and the new method is slower for a fortnight, the manager protects the numbers. That is not obstruction — it is the incentive working exactly as designed.
There is also a legal floor, and it moved this summer. The EU AI Act's literacy duty was rewritten by the Digital Omnibus in July 2026: providers and deployers must now take measures to support the development of AI literacy among their staff, rather than ensure a sufficient level in each person. It is now an obligation of means. That makes it easier to satisfy and no less worth doing properly — because the people asking you for evidence are increasingly clients and insurers rather than regulators, and neither of them has softened anything.
So we work on the four things that actually move usage: role-differentiated capability, manager incentives, an honest answer to the job question, and measurement against a baseline taken before the change. Where no baseline exists, taking one is the first task — without it there is no way to show the programme worked, and no way to show the literacy obligation was met.
How we do it
- 01
Baseline and impact map
1–2 weeks
Who is affected, how their work changes, and where the resistance will be. Plus the before-measurement: current usage, current cycle time, current literacy by role.
- 02
The job question
3–5 days
What leadership is actually prepared to say about roles, and then saying it. This is a leadership decision rather than a communications exercise, and delay is itself read as an answer.
- 03
Role-differentiated capability
3–6 weeks
Executives, managers, the people who operate the system, and staff with incidental exposure — four different needs. Delivered as working sessions on your own use cases, with attendance and assessment recorded so the literacy obligation is evidenced rather than asserted. Where accredited certification is the right answer, that sits in our training catalogue.
- 04
Manager enablement
2–3 weeks
The layer that decides whether this happens. What changes in their targets, how they handle the fortnight in which the new way is slower, and what they do with the first visible mistake.
- 05
Champions and support
ongoing
Named people in each function with hours actually allocated, and a route for the questions nobody will ask in a town hall.
- 06
Measurement and adjustment
from week 4
Usage against baseline, by team. Where a team is flat we find out why rather than repeat the training — the reason is usually a process the system did not remove, or a target nobody changed.
Named artefacts
What you receive
- Impact map by role — what changes, for whom
- Adoption baseline: usage, cycle time and literacy by role, measured before the change
- Leadership position on roles, written and delivered
- Role-differentiated capability programme, delivered
- AI literacy evidence pack — content, attendance, assessment and dates, by role
- Manager enablement pack, including what changes in their targets
- Champion network with named people and allocated time
- Usage reporting against the baseline, by team
- Route for concerns raised outside formal channels
- Handover so the programme continues without us
What we need from you
- A leadership decision on the job question before the programme starts. We cannot answer it for you, and an unanswered one poisons everything that follows.
- Manager time, and a willingness to change what managers are measured on. This is the single biggest predictor of whether usage moves.
- Access to usage data, or agreement to instrument it before launch.
- Champions with hours in their week rather than a title.
- Patience for one quarter. Usage curves do not move in a fortnight, and campaigns that demand it produce compliance rather than adoption.
What changes
- 01Usage moves, and can be shown to have moved, by team, against a baseline.
- 02Your AI literacy obligation has an evidence trail rather than an intention.
- 03Managers are equipped and incentivised rather than merely informed.
- 04The job question is answered by leadership rather than by rumour.
- 05Where a team does not adopt, you learn the reason instead of repeating the training.
What it costs
On request
All prices exclude VAT.
Questions
Can you just deliver the training?
We can, and the accredited options are in the training catalogue. But training alone rarely moves usage — the constraint is usually manager incentives or an unanswered job question, and a course touches neither. If the budget is limited, spend it on the manager layer first.
How does this sit with the EU AI Act literacy obligation?
Since the Digital Omnibus took effect in July 2026, the duty is to take measures that support the development of AI literacy among your staff — an obligation of means, not a guaranteed level in each individual. We deliver by role and record content, attendance, assessment and date. That satisfies the duty comfortably, and more usefully it answers the client and insurer questions that are not going to soften. What your actual exposure is depends on your systems, which the EU AI Act readiness engagement establishes.
Our people are afraid this is about headcount. What do we say?
Whatever is true. If roles will change, say which and by when. If you do not know yet, say that, and say when you will. The one position that reliably fails is reassurance that later turns out to be false — it costs you the next three programmes as well as this one.
Does it have to run alongside a build?
It works best that way, because the change is concrete and there is something real to use. Run ahead of any system it produces enthusiasm with nowhere to go; run after launch it starts from a usage figure you now have to explain.
What does it cost?
On request. It is scoped to headcount, the number of functions affected and how much capability delivery is included. We quote it in two parts, the change programme and the capability delivery, so you can see what you are buying.
Leave with your top three risks documented
Thirty minutes with a senior practitioner. No slideware, no sales engineer.