Service · Diagnostic
Two weeks. Fixed fee. A costed roadmap that tells you which AI opportunities are worth funding, which are traps, and what has to be fixed in your data before either question matters.
Most mid-market companies do not have an AI strategy problem. They have a sequencing problem — several plausible ideas, no shared way of comparing them, and no honest view of whether the underlying data can support any of them.
The assessment exists to end that argument with evidence rather than opinion. It is deliberately small: two weeks, a fixed price, and a scope tight enough that you can approve it without assembling a steering committee. What you get back is yours whether or not we ever work together again.
We sit with the people doing the work — not only the people who manage it — and document where time actually goes. Review queues, rekeying between systems, documents read by hand, decisions that wait on one person's availability. This is where the real opportunities are, and they rarely match the ones on the leadership team's list.
Every candidate is scored on the same four axes: business value, implementation effort, data availability, and risk exposure. The output is a ranked list with the reasoning attached, so your team can re-run the exercise on new ideas after we've gone.
Data quality is where mid-market AI projects most often die quietly. We assess whether the data that a proposed use case depends on actually exists, is accessible, is accurate enough, and can be used lawfully for the purpose you have in mind.
For regulated businesses this is the section that determines whether anything ships. We assess current practice against recognised frameworks — the NIST AI Risk Management Framework and ISO 42001 among them — and identify the specific policies, controls and documentation you would need before a model touches a customer decision.
Who in the organisation can maintain what gets built. Where the internal champions are. Where the scepticism is concentrated and whether it is well-founded. A technically sound system that nobody trusts is a failed project with good documentation.
Every artefact is written in a format your team can read, edit and extend without us. Nothing depends on a proprietary platform we own.
For context, here is what the same diagnostic scope costs across the market:
We quote a specific number once we know the size of the organisation and how many business units are in scope. There is no charge for that conversation.
A 30-minute call is usually enough to tell. If the answer is that you don't, we'll say so.
Book a conversation →Companies roughly between 50 and 1,000 people, or in the $50M to $2B revenue band, where someone has been asked to produce an AI plan and does not want to spend a quarter and a six-figure fee arriving at a slide deck.
It is a particularly good fit if you operate under regulatory supervision — financial services, insurance, healthcare — because the governance work is built into the diagnostic rather than sold as a separate follow-on engagement.
If your requirement is deploying a vendor tool at scale in a stable environment, a certified implementation partner will do it faster and cheaper. If you need a multi-country programme with in-country legal coverage, you need a firm with that bench. We would rather say so at the first call than three weeks into a contract.
Then what
You end with a ranked roadmap and an honest view of what your data and governance can currently support.
One use case into production with the guardrails and evaluation that make its output safe to rely on. See architecture services →
Start here
A short description of the organisation and the problem is enough for us to say whether an assessment is the right next step — or whether something smaller is.