Engagement type
What an AI readiness assessment should actually produce
Last updated
Justas Butkus is a fractional AI officer based in Vilnius, Lithuania, working with mid-market companies and scale-ups across the UK, EU and US. Every engagement begins with a fixed-scope diagnostic, runs on written milestones rather than hours, and leaves accounts and documentation in the client's hands throughout.
Short answer
An AI readiness assessment is a fixed-scope review that answers one question: where can AI pay for itself here, and where can it not. A good one ends with a ranked list, an honest data verdict and a named accountable owner. A bad one ends with a maturity score.
What should an AI readiness assessment contain?
- An inventory of what is already running, sanctioned or not, and what data has passed through it.
- A ranked shortlist of candidate processes, ordered by frequency multiplied by fully loaded cost, not by how interesting they are.
- A data verdict per candidate – not whether data exists, but whether the same field means the same thing everywhere it appears.
- An explicit exclusion list. What you should not automate, and why. This is the section that proves selection actually happened.
- A governance shape proportionate to the risk, including who accepts the risk of going live.
- A first project recommendation that is deliberately unglamorous, to prove the pipeline end to end.
What are the warning signs of a bad assessment?
- A maturity score out of five. It benchmarks you against an average nobody chose and tells you nothing about what to do on Monday.
- No exclusions. If everything you raised is worth doing, no selection took place.
- Confident return figures before any data was examined. Nobody can know that yet, and the number exists to justify the fee.
- A long list of "opportunities". A list is not a decision. Ten unranked possibilities put the prioritising back on you.
- No named owner. If the report does not say who accepts the risk of going live, the pilots will not.
How long should it take?
One to three weeks for a mid-sized company, fixed in scope and with a defined end date. Longer than that usually means the scope was never bounded, and the assessment has quietly become the engagement.
The deliverable is written, and you keep it regardless of whether anything follows. If the honest conclusion is that AI is premature for where you are, that is what it will say – that outcome is worth the fee too, because the alternative is finding out in month nine.
Frequently asked questions
What is an AI readiness assessment?
A fixed-scope review that establishes where AI can and cannot pay for itself in a specific business, ending in a ranked shortlist, a data verdict per candidate, an explicit exclusion list and a named accountable owner.
How long does an AI readiness assessment take?
One to three weeks for a mid-sized company, with a defined end date. If it runs materially longer, the scope was not bounded and the assessment has become the engagement.
What should it deliver?
A written document you keep regardless of what follows: what AI is already running, which processes are worth automating ranked by cost and frequency, which are not and why, whether the data supports them, and who accepts the risk of going live.
Is a maturity score useful?
Rarely. It benchmarks you against an average nobody chose and does not tell you what to do next. A ranked list of your own processes is worth more than a position on someone else's scale.
If this is the piece you need
It is a bounded piece of work with a defined deliverable, which makes it quick to scope.