An AI maturity assessment tells you how far your organisation has got with AI, and where the gaps are. That is all it is. This guide explains what one measures, what the five stages mean, and what to do with the result. It is for organisations of any size and sector.
What is an AI maturity assessment?
It is a structured way of answering one question: how well is AI working for us today?
A good one looks past whether you have bought a tool. It asks who uses AI, what for, how well, and whether anyone is looking after it. Then it places you on a scale, so you can see where you are and what comes next.
The scale is called an AI maturity model. There are many of them. Gartner’s has five levels. MIT’s has four. The names differ, but they all describe the same journey: from a few people trying things out, to AI being part of how the work gets done.
AI maturity or AI readiness: which do you need?
The two get mixed up.
An AI readiness assessment asks whether you can start. Do you have the data, the tools, the skills and the budget?
An AI maturity assessment asks how far you have got. What is in use, how well, and what is missing?
A few years ago, readiness was the right question. Today, your team is almost certainly using AI already, whether you have approved it or not. Our guide to shadow AI shows how much of that leaders never see. So for most organisations, maturity is the more useful question now.
The five stages of AI maturity
These are the five stages we use, with what each looks like from the inside.
- Exploring. AI use is just starting, and it is down to individuals. A few people use ChatGPT or Copilot, often on personal accounts. There are no rules and no plan.
- Experimenting. Some people use AI regularly, but it is not organised. There are pockets of good practice that nobody else knows about.
- Structured. AI use has some shape. There are approved tools, some training, and rules on what data can go in. Someone is responsible for it.
- Scaling. AI is used widely and the basics are in place. The work now is making it consistent across every team, and measuring what it delivers.
- Leading. AI is part of how the organisation works. Ownership is clear, results are measured, and whole processes have been redesigned around it.
A stage is a description, not a grade. Stage 2 is not a failure. It is where many organisations are. And one organisation is rarely at one stage throughout: finance may be at stage 3 while operations is at stage 1.
What an AI maturity assessment measures
Models differ on the detail. Ours scores six areas.
- Adoption. How many of your team use AI, and how often.
- Depth of use. What they use it for. Rewriting an email is one thing. Analysing data or running a whole task is another.
- Collaboration. Whether good ideas spread. Do people share prompts and assistants, or does everyone work it out alone?
- Skills and enablement. Whether people have had training, and whether they know how to check an answer.
- Outcomes. Whether AI is saving time or improving the work, and whether anyone measures it.
- Governance and safety. Whether there are approved tools, clear rules on data, and a named owner. Our guide to AI governance covers this one in full.
Each area gets a score from 0 to 100. The overall score is the average, and that sets the stage.
These six are a summary of what we look at, not the full method. Our free AI Maturity Assessment covers them and more.
Why one person’s score is not enough
A common way to assess AI maturity is a short quiz for one leader: you answer questions about your organisation and get a score.
The trouble is that leaders see the official picture: the tools the company bought, the policy that was sent round. Your team sees the real one: what they use day to day, what they avoid, and what they have never been shown.
So a single score tells you what one person believes. It also hides the differences between departments, which is usually where the useful answers are.
The better way is to ask your team, anonymously, so they answer honestly. That is how our assessment works.
What a good report tells you
A stage number alone is not much use. A good report should give you:
- your stage, and your score in each area
- what is going well, so you can build on it
- the risks to deal with first
- the differences between departments
- a short list of priorities
You can see what ours looks like in the sample report.
What to do in the first 30 days after your results
- Week 1. Share the results with your leadership team. Agree the two or three priorities that matter most.
- Week 2. Deal with the biggest risk first. That is usually data going into personal accounts.
- Week 3. Find the team that is furthest ahead. Ask them to show the others what works.
- Week 4. Give each priority a named owner and a date.
Then run the assessment again once the first priorities are in place, to see what has moved.
If you want help with what comes next, see how we help. If the priority is getting Copilot, ChatGPT or Claude used well by everyone, see our AI rollout page. And if nobody obvious can own the plan, a fractional Chief AI Officer can lead it part-time.



