FRA9 · 04/09/2026 · Digital & artificial intelligence
France’s Osez l’IA 2026: SME diagnosis, cost and support
The programme is moving from awareness to deployment: who can request a Diag Data IA, what it costs, what it delivers and what it does not fund.
The programme is moving from awareness to deployment: who can request a Diag Data IA, what it costs, what it delivers and what it does not fund.
The results published on 3 September
One year after Osez l’IA launched, France’s Economy Ministry reports more than 35,000 businesses reached, 615 AI Ambassadors across 20 regions and authorities, and 13 sectors covered. The national catalogue lists 88 AI solution providers for SMEs and mid-caps, while the Accélérez avec l’IA platform contains 54 use cases. These figures measure awareness and support infrastructure; they do not mean that 35,000 grants or completed AI deployments have been delivered.
Who can request a Diag Data IA
The current offer is for SMEs and mid-sized companies with 10 to 2,000 employees that are registered in France’s trade and companies register when applying. It suits a business with real processes and data but no settled first use case. A microbusiness below the ten-employee floor, an idea-stage founder or an unregistered organisation should begin with France Num resources, its chamber of commerce, a local AI Ambassador or another suitable funding route.
The real price: €6,000 before tax
The diagnosis provides eight consultant-days. France 2030 covers 40% of the €10,000 pre-tax cost, leaving €6,000 before tax for the beneficiary company. This is neither a cash grant nor a flexible reimbursement for software, servers or payroll. The company buys a structured service from a Bpifrance-selected expert and should budget for VAT, internal staff time and any implementation costs that follow the diagnosis.
What the eight days should deliver
The expert reviews technical and operational maturity, identifies use cases, helps prioritise them, estimates cost, possible return and organisational impact, then produces a tailored action plan. Before work starts, appoint an internal lead and prepare process maps, data sources, business pain points, GDPR and security constraints, and baseline performance measures. Without access to operational teams and usable data, even a sound method can produce a generic deliverable.
From 70 diagnoses under way to a target of 1,000
The 3 September release says 70 diagnoses under the current format are already under way and announces the deployment of 1,000 Diag Data IA assignments in the programme’s next phase. France Num also says 2,000 assignments are co-funded for 2026–27 and that earlier formats have already served more than 1,200 companies. Those scopes should not be added as one cohort; applicants must check current availability on Bpifrance’s platform.
Other ways into the programme
A company not ready for a paid diagnosis can use the online AI self-assessment, free training, a Café IA, chamber-of-commerce support, France Num advisers or local AI Ambassadors. More mature firms can examine the 18-month AI Accelerators, co-funded at 46%, regional schemes and European Digital Innovation Hubs. Support helps choose and prepare a project; it does not itself replace a product budget, R&D grant or equity investment.
Five checks before applying
Confirm the 10–2,000 employee range and French registration, then appoint an executive sponsor. Quantify two or three business problems, confirm that the data can be used lawfully, and reserve €6,000 before tax plus internal time. Apply through Bpifrance’s Diag Data IA platform or contact a Bpifrance account manager. Do not commit to a technology deployment before comparing the action plan, recurring costs, security requirements and the public support that can legally be combined.
Read next
