Key points in 30 seconds
- AI maturity measures an organisation's ability to create value from AI, not the number of tools it has bought.
- It is assessed across five dimensions: data, processes, tools, skills and governance.
- Four levels are enough to position yourself: exploratory, experimental, operational, integrated.
- The output of the assessment is a roadmap with three priority use cases.
Why measure maturity before investing
Most AI projects that fail in SMEs do not fail because of the technology. They fail because the data is scattered, the target process is not documented or no one has been appointed to own the tool. A maturity assessment reveals these obstacles before you spend.
The five dimensions to assess
1. Data
Where is your customer, sales and financial data? Is it centralised, up to date and structured? A company whose sales history lives in inboxes and personal Excel files will first need to consolidate it, often in a CRM.
2. Processes
AI automates what has been described. Are your key processes (handling a request, follow-ups, invoicing) documented? Do you know how long they take? A process that everyone runs their own way must be standardised before it is automated.
3. Tools
Do your software systems talk to each other? Do they offer APIs or connectors? An open ecosystem lets you add AI piece by piece; closed software often forces a costly workaround.
4. Skills
Do your teams already use AI assistants, even informally? Can they write a precise prompt and check the output? Training is often the lever with the best cost-to-impact ratio.
5. Governance
Who decides which uses are allowed? Which data may leave the company? Personal data regulations frame these choices. A simple one-page usage policy protects the company and reassures teams.
The four-level framework
| Level | Typical situation | Priority |
|---|---|---|
| 1. Exploratory | Individual, informal use of consumer tools, scattered data | Centralise data, set usage rules |
| 2. Experimental | One or two tests run by volunteers, without measurement | Choose a use case and measure a baseline |
| 3. Operational | One automated process in production, ROI tracked | Scale up and extend to other teams |
| 4. Integrated | AI is part of processes and decisions, governance in place | Optimise, innovate, develop new offers |
How to run the assessment in practice
- Interviews with leadership and two or three operational managers (45 to 90 minutes in total).
- Mapping of the most time-consuming processes and the tools in use.
- Scoring each dimension from 1 to 4, with justification.
- Selection of three priority use cases based on impact and feasibility.
- Roadmap over 3, 6 and 12 months.
The VSIA AI audit follows this method: a 45-minute interview, then a report within 48 hours with your maturity level and your three priority use cases. It is free, with no commitment.
Common mistakes
- Confusing AI maturity with software equipment.
- Running the assessment with the IT team only, without the business teams.
- Aiming straight for level 4 without consolidating the data first.
Example audit result
| Dimension | Level observed | Priority action |
|---|---|---|
| Strategy | 1: no written AI objective | Define two use cases linked to margin or time |
| Data | 2: client data scattered | Centralise in a CRM |
| Tools | 2: individual use of consumer tools | Choose a business plan |
| Skills | 1: no training | Train the pilot teams |
| Governance | 1: no rules | Write a one-page charter |
This kind of summary lets you set priorities in a single meeting.
Questions to ask teams
- Which tasks do you repeat identically every week?
- Where do you copy information from one tool to another?
- Which client questions come up most often?
- Do you already use AI tools, which ones and for what?
- What worries you about the arrival of AI?
The answers often reveal existing uses and the obstacles to remove.
After the audit: the roadmap
The audit leads to two or three priority projects, chosen for their measurable gain and feasibility. The first project should be short, a few weeks, to prove value and build support. The next ones build on the data and skills developed. See our guide on AI use cases with fast ROI.
Repeat the audit every year
Maturity changes quickly, as do the tools. Repeat the assessment every year with the same grid to measure progress on each dimension. This also helps justify investments to management. Our free 45-minute AI audit gives a first reading.
In summary
An AI maturity audit assesses five dimensions, places the company on a four-level grid and leads to two or three priority projects. It avoids investing in tools before having the necessary data, skills and rules. Repeated every year, it measures progress and guides further investment.
Frequently asked questions
What is a company's AI maturity?
It is its ability to use artificial intelligence in a useful and controlled way. It depends on data quality, process documentation, tool openness, team skills and governance.
How long does an AI maturity assessment take?
For an SME, an initial assessment takes one or two interviews of 45 to 90 minutes, followed by a report. At VSIA, the report is delivered within 48 hours.
Do you need an IT department to launch an AI project?
No. Many AI solutions for SMEs can be deployed without an IT department. You do, however, need a business owner to lead the project and accessible data.
How long does an AI maturity audit take?
A first diagnosis fits into a 45-minute interview. A full audit, with interviews and process analysis, usually takes one to three weeks.
Do you need perfect data to start with AI?
No. Many use cases work with existing documents and emails. The audit shows which data to improve first.