What an AI readiness assessment looks like (and the report you should demand)

An AI readiness assessment is a short, structured audit that tells you whether AI can pay off in your business, where, and at what cost, before you commit to building anything. A good one looks at four things: the state of your data, the shape of your day-to-day processes, a shortlist of concrete use cases ranked by value and effort, and a costed plan for the one or two worth doing first. It usually runs two to four weeks. The deliverable is not a strategy deck full of ambition. It is a document you can act on: what to build, why it is worth it, what it will cost to build and to run, and how you will know it worked. If the report cannot be turned into a decision, the assessment has not done its job.

What does an AI readiness assessment actually cover?

It covers the gap between where you are and where a working AI system could take you. Four areas, in order.

First, data. Where does it live, who owns it, is it clean enough to use, and are you allowed to use it the way you intend? Most AI projects stall here, so this comes first.

Second, processes. The assessor maps how work moves through your business today: the manual steps, the handoffs, the places people retype the same figures or wait on someone else. This is where the realistic opportunities hide.

Third, a use-case shortlist. Every candidate is scored on the value it would create against the effort to build and run it. Most ideas should be set aside. The point is to find the one or two with the best ratio.

Fourth, the costed plan. The top candidates get a build estimate, a running cost, and a clear statement of the outcome you are buying.

How does a readiness assessment work, step by step?

The good ones follow a predictable path. You should be able to see each step happening.

1. Scoping conversation. You describe the problems that bother you. The assessor listens more than they pitch. The output is an agreed list of areas to examine.
2. Data review. They look at the systems and records behind those areas: quality, access, ownership, and any legal constraints on use.
3. Process mapping. They sit with the people doing the work, or watch the steps directly, to see how a task really runs rather than how the org chart says it runs.
4. Use-case generation and scoring. Candidates are written down and ranked by value against effort, so the shortlist is defensible rather than a matter of taste.
5. Costing and planning. The top one or two get a build cost, a running cost, and a measurable success criterion.
6. Report and walkthrough. You get the written document and a session to challenge it. If you cannot poke holes in it, it was not detailed enough.

What should a good readiness assessment report contain?

This is the part to be strict about. A report you can act on contains all of the following.

Section	What it must answer	Why it matters
Data findings	Is your data fit to use, and what needs fixing first?	Decides whether anything is buildable at all
Process map	How does the targeted work run today?	Shows where time and money actually leak
Use-case shortlist	Which two or three ideas are worth it, ranked?	Stops you funding the wrong thing
Costed plan	Build cost, running cost, and timeline for the top pick	Turns ambition into a decision
Success measure	How will you know it worked?	Makes the result checkable, not a matter of opinion
Risks and constraints	What could stop this, including legal and data limits	Surfaces the blockers early, loudly

If a report skips the costs, hand-waves the running expense, or lists ten exciting ideas without ranking them, push back. A shortlist of everything is a shortlist of nothing.

What should you demand before you sign anything?

Demand a number and a measure. For the use case the assessor recommends first, ask what it costs to build, what it costs to run for a year, and the single metric that will tell you it worked. If they cannot give you all three, the work is not finished.

Demand honesty about the data. A trustworthy assessor will sometimes tell you the answer is not AI yet, because the data needs cleaning or the process needs fixing first. That is a sign they are looking at your business rather than selling you a project.

Demand that the strategy and the build live in the same place. An assessment written by people who will never have to ship the system tends to drift into the comfortable and the vague. When the team that scopes the work is also the team that could build it, the plan stays grounded in what is actually possible. This is one reason AI consulting in Cyprus increasingly bundles the audit with the capacity to deliver, rather than handing over a deck and walking away.

The practical takeaway

A readiness assessment is worth paying for when it ends with a decision rather than a discussion. Before you commission one, agree on the deliverable in writing: a costed plan for one or two ranked use cases, with a success measure you can check later. If a provider will only promise a strategy document, you are buying advice. If they will commit to a costed, measurable first step, you are buying a starting point you can actually build on. When you have a problem you suspect AI could solve, tell us the problem and we will tell you, plainly, whether an assessment is the right next move; readiness assessments are the first step of our own AI consulting engagements.