AI maturity of an SME: the six axes to assess and the most frequent mistakes
AI maturity is not measured by the number of tools installed. Six axes, four levels, and the two imbalances found in most SMEs.

Article prepared with the help of artificial intelligence from our monitoring, then reviewed and approved by Christophe Delebarre before publication.
A company that has taken out three subscriptions to AI assistants is no more mature than one that has none. It just has more tools. Maturity is the ability to draw measured value from AI, repeatably, without endangering its data or its teams. It is judged on several dimensions at once, and it is the imbalance between them that explains most failures.
The six axes
The AI Radar, our twelve-question self-assessment, measures six axes. They are the ones we find in every diagnostic.
- Strategy and steering. Does AI answer identified business objectives, with indicators defined before starting, or is it a series of individual experiments? Who decides, who arbitrates?
- Data. Is the data your processes depend on held in accessible, up-to-date, reliable tools? You do not automate what rests on a spreadsheet with six versions.
- Technology and integration. Do your tools have interfaces that let an assistant read and write: ERP, mailbox, calendar, files? A closed tool is a wall.
- Compliance and trust. Do you know where your data goes, with which processor, under which contract? Do you have a usage framework, human validation on the actions that matter, information for the people concerned?
- Teams and culture. Are your staff trained, involved in choosing use cases, free to say what does not work? The AI literacy obligation of the European regulation makes it a regulatory axis as much as a cultural one.
- Operations. Once in place, who monitors results, who maintains, who decides to extend or stop? An automation without an owner degrades silently.
The four levels
On each axis, four levels: discovery (the topic is known, nothing is committed), initiation (individual experiments, no framework), structuring (rules, owners, first measured automations), maturity (AI is part of normal operations, with steering and continuous improvement). Most SMEs we meet are between initiation and structuring, with large gaps from one axis to another. That is normal, and it is precisely what the profile reveals.
The two most frequent imbalances
Technology ahead, compliance behind. The company has tools, experiments, sometimes a first assistant in production, but nobody knows where the data goes, no usage framework is written, teams use consumer tools with customer data. It is the most common profile, and the riskiest: the day a customer or the data protection authority asks the question, there is no answer. The remedy costs almost nothing: a one-page usage framework, team training, a choice of tools under contract.
Strategy ahead, data behind. Management has a vision, objectives, sometimes a budget, but the targeted processes rest on scattered or non-existent data. Projects start, then bog down in data preparation. The remedy is to pick a first process whose data is already clean, even if it is less strategic, to learn while winning.
The mistakes that keep coming back
- Confusing maturity with equipment. Buying licences does not raise the "teams" axis or the "compliance" axis.
- Assessing alone. An owner who answers the twelve questions without their teams almost always overestimates the "data" and "teams" axes. Have two or three people answer and compare.
- Aiming for maturity everywhere. A twenty-person SME does not need an AI ethics committee. It needs a simple framework, rules that are applied and a first automation that holds.
- Not repeating the assessment. A profile six months apart shows whether the approach is moving. That is its real use.
What next
The AI Radar takes five minutes and emails you your profile. It does not replace a diagnostic, it tells you which axis to start with. To go further, the AI Diagnostic maps your processes and data with your teams and leads to a roadmap with the first prototypes and their quantified criteria. And to place your company on these six axes in a few minutes, take the AI Radar.
