AI Opportunity Scan
Where AI helps you,
and in what order.
Interviews across the business, then a prioritised roadmap of use cases with effort and value against each. It ends on one recommended first build with a rough cost, so what you get is a decision rather than a document about decisions.
The problem
A list of things AI could do for you is not useful. The order is.
Any competent consultant can produce forty ideas in a fortnight. The work that matters is cutting them to the two or three worth doing, in a sequence where each one funds or de-risks the next, and being specific enough about the first that somebody could start on Monday.
That needs to know how the business actually runs, which is why this starts with interviews rather than a workshop. The people doing the work know where the waste is. They have usually never been asked in a way that made it safe to say so.
What you get
Interviews across the business
Not only the people who commissioned the scan. The waste is rarely visible from the top of an organisation.
A prioritised roadmap
Use cases in a defensible order, with what each depends on, rather than a matrix that leaves the sequencing to you.
Effort against value
An estimate for each, honest about which numbers are firm and which are a guess with a reason attached.
One recommended first build
Named, scoped and roughly costed. The point of the exercise is that you leave with something to decide about.
How it runs
Two weeks, remote, with a working session at the end rather than a document dropped in an inbox.
Interviews
Six to twelve conversations across functions and seniority, an hour each. Where the work is, where it stalls, and what people already work around.
Shape and score
Candidate use cases, scored on effort and value, and cut hard. The cutting is most of the value.
The first build
The recommended one, scoped properly, with a rough cost and the data conditions it depends on.
The session
A working session to walk it through with the people who will fund it, so the questions get answered live.
What sits behind it
The scan exists because the build usually follows it.
- Prima PolisThe kind of thing a scan points at: four vendor systems made to cooperate so advice records write themselves, under Dutch financial regulation.
- ConsumrzPrioritising across three European markets from market, competitive and financial analysis, then committing to a sequence.
- Tata Steel, GRC in ServiceNowEstablishing how data actually moved across an enterprise, which is the question most AI roadmaps quietly skip.
What it costs
EUR 9,500
Fixed fee, two weeks, remote
Fixed, published, and the same for everyone.
Two weeks of work at a fixed price, so you can judge the size of it before we speak. There is no discovery phase that discovers the price is higher than quoted.
Where the scan concludes that the honest answer is to do nothing with AI for now, it says that. It has happened, and it is worth the fee.
Fit
Best when you can act on the answer.
This is worth doing when there is budget and appetite to build something afterwards. A scan commissioned to satisfy a board that asked about AI produces a document nobody acts on, and you would be better off with the briefing.
If your pressure is regulatory rather than commercial, the position and roadmap is the one that answers it.
Common questions
- How many of our people does this take?
- Six to twelve interviews of about an hour, plus the closing session. Chosen across functions and levels rather than only the sponsors.
- Do you need access to our systems?
- No. This is interviews and documents. Access becomes relevant if a build follows.
- What if we already know what we want to build?
- Then say so and we can scope the build directly. A scan that confirms a decision already made is an expensive way to feel reassured.
- Does it cover the AI Act?
- Lightly, where it touches a use case. If the regulation is the actual question, the position and roadmap is the right engagement.
Worth doing when you know AI should be useful here and cannot yet say where.
Start a conversation