We've started, but I'm not sure we're heading in the right direction
For you if you have bought licences, run pilots or built something, and cannot yet say whether it is paying back, or why it has stalled. You are not alone: CMI's 2026 research found 68% of UK managers “still in the early stages of experimenting or building pilots”.
In 20 minutes we can cover
- what you set out to change, and how you would show your board the return. The usual measure, the same work done more cheaply, misses the rest: work you could never have done before, and what the firm could now be doing;
- whether you are checking too little (errors reach clients) or too much (the time AI saved goes on rework), and whether your juniors are still learning to spot work that looks right and isn't;
- if you are regulated: what your regulator and your insurer will expect to see about how AI work is supervised.
You'll come away with
A straight view of whether what you are doing is worth continuing, what to look at first, and what to measure to show the return. If it needs a proper look, the next step is a Diagnose (below).
Worth a look first
Book 20 minutes
Pick a time that suits you, or simply reply to my email.
You'll talk to me
Richard O'Rourke: engineer, three companies founded and run as chief executive, and now research on how AI work should be supervised. What you tell me stays confidential. More about me.
If it's worth going further: the Diagnose
A fixed-scope, three-week review of the AI already in use in your firm: where it is used, what it draws on, who checks what before it leaves, and where the checking load sits. It ends in written priorities, presented to your leadership. Fixed fee: £9,000.