The point is to make ourselves unnecessary
Most AI training is a recorded webinar about what a large language model is. It produces awareness and no change in behaviour. What actually shifts an organisation is role-specific practice on real work, clear rules people can follow under pressure, and a route for good ideas to be built rather than discussed.
- Measured
- capability baseline and uplift, rather than a satisfaction survey
- Every role
- receives concrete rules on what must never be entered into a tool
- Quarterly
- cadence on the front-line idea intake, with a stated response
- Falling
- reliance on external support as internal capability matures
What it looks like today
The pattern we find in almost every operation
- Staff paste customer data, contracts and source code into public AI tools with no guidance and no record.
- A one-off training webinar was delivered, attendance was recorded, and nothing changed.
- Front-line staff have good automation ideas and no route to have them evaluated.
- Teams that do use AI well keep it to themselves, so the benefit does not spread.
- Purchased tooling sits at low adoption because nobody was shown where it fits their actual job.
- The organisation cannot tell whether any investment in training worked, because capability was never measured.
How we do it
The approach, step by step
- 01
Measure the starting point
A baseline of what each role can currently do, what tools they are already using silently, and where the real risk is. Without this, any later claim of improvement is an anecdote, and you cannot tell which teams need which intervention.
- 02
Teach the job, not the technology
Role-specific sessions on the work people actually do — reviewing a contract, drafting a customer response, triaging a ticket, interrogating a report — with the tooling they already have access to. Generic education is cheap to deliver and reliably produces no behaviour change.
- 03
Give rules people can follow under pressure
An acceptable-use policy written as concrete decisions rather than principles: what must never be entered, what must always be verified, what must be disclosed. People follow rules they can apply while busy, and ignore rules that require interpretation.
- 04
Open a route from the front line to production
A lightweight intake for AI and automation ideas, with a stated triage and response, so the people closest to the work can contribute to it. This is where the genuinely valuable use cases come from, and the mechanism has to survive the first three ideas being rejected.
- 05
Build a community with named champions
Champions in each function, given time and standing, sharing what worked and what did not. Sustained capability comes from peers rather than from an external programme that ends on a date.
- 06
Re-measure so the improvement is provable
The same baseline instrument repeated after ninety days and six months, reported to the sponsor. It shows where the investment worked, where it did not, and it is the only way to justify continuing to fund capability rather than tooling.
What you receive
Deliverables, stated up front
- Capability baseline per role, with the risk of current practice documented
- Role-specific curriculum built from real tasks, not from AI theory
- Acceptable-use policy expressed as concrete, checkable rules
- Delivered training with competence checks rather than attendance records
- Intake and triage process for front-line AI and automation ideas
- Champion network with named people, allocated time and standing
- Internal knowledge base of working patterns and known failure modes
- Ninety-day and six-month capability re-measurement against baseline
- Capability report to the sponsor with recommendations for the next cycle
You are a fit if
- You have bought AI tooling and cannot demonstrate what it returned
- You suspect staff are using public AI tools with data they should not
- Training has been delivered and behaviour has not changed
- Front-line staff have ideas and no route to get them assessed
- You want to reduce long-term reliance on external providers
We will tell you it is a fit problem if
- You want a one-off training day and a completion report, which will not change anything
- The organisation has not decided what employees are permitted to put into AI tools, and is unwilling to
- You are measuring success by attendance rather than by capability
Systems we work with
Not on the list? We integrate against anything with an API, a database, a file interface or a documented import format.
Questions we get on this
AI literacy: the practical answers
How is this different from what our learning and development team already does?
Isn't prompting about to become irrelevant?
How do you know capability improved?
Do you push a specific vendor's tools?
Bring us the process you already know is costing too much
Thirty minutes with an engineer is usually enough to tell whether it is worth automating, roughly what it would save, and whether the payback is inside a window your finance team will accept. If the answer is no, we will say so on the call.