Entry engagement
AI readiness assessment
A fixed-scope entry engagement, two to four weeks, at a price agreed before it starts. It answers two questions: which specific workflows would repay AI investment now, and what would have to be true for the first one to reach production. It does not produce a maturity score.
The business problem
Readiness assessments are worth doing and easy to do badly. The recognisable failure is eight weeks of interviews, a five-level maturity model, a score somewhere in the middle, and no decision anyone can act on.
The UK government's AI adoption research found that among organisations already using AI, only about half felt ready to scale it. Readiness is not a precondition for starting. It is what most organisations discover they lack after they have already started.
This engagement is the short, cheap version of that discovery. Our full view of what the exercise should contain is in what an AI readiness assessment should include.
Who it is for
Organisations with more candidate use cases than capacity, which is almost all of them, and anyone who has been asked to "do something with AI" and wants to arrive at a defensible answer rather than a list of pilots.
It also suits the opposite situation: a decision that already feels made, where an independent read on the data and the workflow is cheaper than discovering the problem in month four.
It is not for you if the build is already funded and you want validation rather than assessment. We would rather say that now than take the work.
What is assessed
Six dimensions. We weight them to your situation, but we do not skip any, because the weakest one determines the outcome.
Candidate workflows
Specific decisions with a monthly volume, a current cost or cycle time and a known error rate. Not themes like customer service.
Data reality
Where the inputs live, who owns them, what state they are actually in. We ask to see samples of the real records.
Business context
The rules, exceptions and precedents a competent new joiner would need. If it exists only in people's heads, that is a finding.
Ownership and authority
Who is accountable for the number today, and whether they can change the process around it.
Governance posture
Where you sit against the NIST AI Risk Management Framework and what the EU AI Act implies for these use cases.
Operating capability
Whether the organisation can run a service to a quality target and change it safely. Existing software maturity predicts AI outcomes better than any AI-specific measure.
How the assessment works
Two to four weeks. The calendar constraint is data access and people's diaries, both of which can be arranged in advance.
What you will receive
Four things, and deliberately nothing else. If it runs past twenty pages the shortlist is buried in it.
- A ranked shortlist. Three to five candidate workflows with volume, current cost, expected effect and the main uncertainty for each.
- One recommended first build. With the number it should move, the threshold that would justify scaling, the named owner and the condition under which you stop.
- A blockers list with owners and dates. Data access, approval paths, missing documentation. Each one assigned, because a blocker without a name is a note.
- An explicit not-now list. What we considered and rejected, with reasons. This is the page clients reuse most, because the same ideas return every quarter.
What the assessment enables
The decisions you are able to make afterwards that you could not make before.
Start or do not start
A defensible answer either way, reached in a fortnight rather than a quarter.
Where to spend first
One workflow chosen on volume, cost and feasibility rather than on which demo was most convincing.
What to fix before building
A named list of blockers, so the sequence is realistic rather than optimistic.
What to decline
A written basis for saying no to the ideas that return every quarter.
Whether to build or buy
Enough evidence to take the layer-by-layer decision rather than a single verdict.
Who owns it
An accountable person identified before the work starts, which is the most common missing piece.
Frequently asked questions
How long does it take?
Two to four weeks. Longer than that and the assessment becomes the project it was meant to scope.
What does it cost?
A fixed price, quoted before we start, because the scope is fixed. Ask us and we will tell you on the first call.
Do we need clean data first?
No. Establishing what state the data is actually in is part of the assessment, and it is usually different from what people believe.
What if the answer is that we are not ready?
That is a common and useful outcome. The deliverable then names the specific blockers with owners, which is a cheaper finding here than in month four of a build.
Are you assessing us or selling us a build?
The not-now list is a required deliverable. A firm that never recommends against building has not given you an assessment, and we hold no reseller agreements either way.
Who needs to be involved?
The people who do the work today, whoever owns the numbers, and someone who can speak for data access and security. A few hours each.
Do you produce a maturity score?
No. A score grades you; a shortlist decides something. We produce the shortlist.
What happens afterwards?
You hold a recommendation you can act on with anyone. If you want us to build it, that is a separate engagement with its own decision point.
Relevant insights
Start a conversation.
Fixed scope, fixed price, two to four weeks. If the honest answer is that nothing clears the threshold yet, you will hear it early and in writing.