Bpifrance's Diag Data IA part-funds 8 days of work by an approved data and AI expert, spread over 3 months at most, to assess AI's potential in your company, identify use cases and prioritize them in a roadmap. It is aimed at French SMEs and mid-sized companies with 10 to 2,000 employees and at least €1 million in revenue. The quote is capped at €10,000 excluding VAT, of which €6,000 excluding VAT remains to be paid after a 40% subsidy.
What is Bpifrance's Diag Data IA?
The Diag Data Intelligence Artificielle is a Bpifrance program that helps a company identify value-creating projects from its own data. It is carried out by a data science and artificial intelligence expert approved by Bpifrance. According to Bpifrance, it is open whatever your maturity with AI and data, and particularly targets companies that want to innovate and grow through AI projects.
The offer is funded with support from France 2030, as part of the "Osez l'IA" plan.
Which companies are eligible?
According to Bpifrance's official page, the diagnostic is open to companies that meet these conditions:
- An independent SME or mid-sized company with 10 to 2,000 full-time equivalent employees, whether or not a Bpifrance client
- At least €1,000,000 in revenue over a 12-month financial year
- More than one year in business
- Any sector eligible under Bpifrance's policy, located in France, overseas territories included
- Excluding companies in difficulty as defined by EU regulation
How does the Diag Data IA work?
The diagnostic covers 8 days of the expert's time, spread over 3 months at most, for three deliverables:
- A technical and operational review of how AI could fit into the company
- Concrete use cases suited to your context and ambitions
- Those use cases prioritized by the value they add, and a roadmap
How much does the Diag Data IA cost?
The quote is capped at €10,000 excluding VAT. For companies that meet the EU definition of an SME and for mid-sized companies with fewer than 2,000 employees, Bpifrance subsidizes 40%: €6,000 excluding VAT remains to be paid.
These are the amounts on the official page when we read it. Check them with Bpifrance before committing.
How should you prepare to get the most out of it?
Eight days go quickly. Have your material ready before it starts, so the expert spends their time on analysis rather than collection.
- The repetitive tasks that take the most time, team by team
- The tools you use and the data they hold (CRM, ERP, spreadsheets, documents)
- One contact person per team, available for interviews
- Your constraints: personal data, security, budget, timing
What should you do after the diagnostic?
The roadmap only helps if a first use case gets started. Pick the one that combines real value with a simple implementation, test it on real cases, then move on to the next.
If you would like a quick opinion first, before committing to a multi-day diagnostic, a short first conversation can help you decide whether AI has a place in your business and where to start.
Key takeaways
- 8 days of an approved expert over 3 months at most, for prioritized use cases and a roadmap
- SMEs and mid-sized companies with 10 to 2,000 employees and at least €1 million in revenue
- Quote capped at €10,000 excluding VAT, €6,000 left to pay after a 40% subsidy
- Prepare your tasks, tools and data before it starts
FAQ
Is the Diag Data IA free?
No. According to Bpifrance, it is 40% subsidized for eligible companies: on a quote capped at €10,000 excluding VAT, €6,000 excluding VAT remains to be paid.
Can a very small business get the Diag Data IA?
The official page targets SMEs and mid-sized companies with 10 to 2,000 employees and at least €1 million in revenue. A smaller company falls outside these criteria.
Who carries out the diagnostic?
A data science and artificial intelligence expert approved by Bpifrance.
How is it different from a free AI audit offered by a provider?
The Diag Data IA is a co-funded, multi-day program that ends with a roadmap. A short audit, such as a 45-minute first conversation, is more about knowing where to start and whether such a program is worth it.
Sources
- Bpifrance: Diag Data IA (accessed September 29, 2026)