What Does an AI Readiness Assessment Actually Tell You?
It should tell you two things: which use case to do first, and what would have to be true for it to work. A good assessment is short, specific and occasionally disappointing — if your data cannot support the thing you were hoping to build, finding that out in two weeks is the cheapest result available to you.
A Vague Assessment vs. a Useful One
| Step | Vague Assessment | Useful Assessment |
|---|---|---|
| The output | A maturity score and a slide on trends | A ranked shortlist with the first project scoped |
| Data review | Asks whether you have data | Inspects the actual tables you would need to use |
| Feasibility | Asserted from industry benchmarks | Tested against your records, volumes and quality |
| Bad news | Avoided, because it complicates the sale | Stated plainly, with what would have to change |
| What you can do next | Commission a larger engagement | Start the first build, or fix a named prerequisite |
Three Questions That Decide Everything
Is the work repetitive enough that a pattern exists? One-off judgements with no history behind them are the hardest thing to automate and the most commonly proposed.
Can someone define a good outcome? If two experienced people in your team would disagree about whether a given result was correct, a model cannot learn the distinction and you cannot measure it afterwards.
Do you have the data, in a system, at sufficient volume? Not in principle — actually, in a table someone can query. A surprising number of projects stall here, and it is far cheaper to discover it during an assessment than during a build.
What You Should Get at the End
A shortlist of use cases ranked by value and feasibility, with the reasoning shown for each — including the ones ruled out and why, because that saves you from being sold them later.
For the top candidate: what data it needs, what a first slice looks like, what would have to be true for it to work, and the numbers to baseline before starting. If an assessment does not tell you how you would know it failed, it has not finished its job.
Where This Fits
This is one part of our work in AI for Strategic Management. See the full set of AI use cases for the equivalent in other industries and functions.
Frequently Asked Questions
How long should an assessment take?
Two to four weeks for most organisations. Shorter than that and nobody has looked at real data; longer and it becomes a project in its own right, which defeats the purpose. The point is a decision, not a document — if the assessment costs a meaningful fraction of the build it was supposed to de-risk, something has gone wrong.
What if the assessment says we are not ready?
Then it just saved you considerably more than it cost. 'Not ready' should always come with specifics — which data is missing, which process needs defining, what would change the answer. A verdict without a route forward is not an assessment, it is an opinion.
Do we need a data warehouse first?
Usually not, and this is where a lot of budget goes to die. Plenty of first AI projects run on data extracted from operational systems, and a full warehouse programme can absorb a year before anyone sees a result. Build the warehouse because your reporting needs it, not as a prerequisite someone told you was mandatory.
Who needs to be involved from our side?
Someone who knows the process end to end, someone who knows where the data actually lives, and someone who can decide. The third is the one most often missing, and its absence is why assessments turn into circulated documents rather than decisions.
How is this different from a strategy engagement?
Scope and output. A strategy engagement produces a multi-year view; an assessment produces a first project you could start next month. Both have their place, but if you have never shipped an AI system, the assessment is the one that teaches you something real — and it makes the eventual strategy far better informed.
Find out in weeks whether your first use case is buildable.
Get a Free Proof of Concept within weeks.
Let’s talk
AI is here to stay
Let’s win together
Your first 45-min alignment session — and a small PoC — are free.
Or just say hello or write us an email.