Osez l’IA turns one: what SMEs and mid-sized companies can do in practice now

On 3 September 2026, Anne Le Hénanff, French Minister Delegate for Artificial Intelligence and Digital Affairs, brings together the stakeholders of the Osez l’IA plan at the Ministry of the Economy (Bercy) for its first anniversary. The event, organised by the Directorate-General for Enterprise (DGE), presents the one-year review and the outlook for the year ahead.

For a European SME or mid-sized company, this is not a ceremony to watch from afar. It is a reminder: the French state has put in place a framework, local contacts and co-funded programmes. The useful question is no longer “should we look into AI?” but “where do we start without spreading the budget and the timeline too thin?”.

What the plan really aims for

Launched in July 2025 as part of the national AI strategy, Osez l’IA aims, by 2030, for an AI adoption rate of 100% in large companies, 80% in SMEs and mid-sized companies, and 50% in very small businesses.

The plan rests on three levers: awareness, training, and support to identify relevant solutions and take action. In other words: fewer slides, more prioritised use cases.

One year later: the official figures

According to the plan’s official page (updated on 30 July 2026) and the DGE announcements of June 2026:

• 615 AI ambassadors across 20 regions and local authorities (overseas territories included) and 13 business sectors;
• more than 35,000 companies reached by awareness actions in nine months;
• an official catalogue of 88 companies offering AI solutions suited to SMEs and mid-sized companies (DGE and Hub France IA);
• an “Accélérez avec l’IA” platform (DGE and Bpifrance) with concrete deployment cases.

These figures show real mobilisation. On their own, they do not say that a given company already has a use that lasts. Between awareness and the first industrialised process, there is still the diagnostic, prioritisation, team training and budget decisions.

Two funded programmes to use (if you are eligible)

1. Diagnostics Data IA (Bpifrance Conseil, co-funded by France 2030)

• An 8 person-day engagement;
• cost of €10,000 excl. VAT, 40% covered by France 2030;
• goal: a technical and operational assessment, identification of use cases, prioritisation by added value;
• eligibility: SMEs and mid-sized companies with 10 to 2,000 employees, registered in France on the trade and companies register (RCS).

2. Accélérateurs IA (18-month support programme)

• A 360° diagnostic with an AI focus, strategy modules, deployment support, group training and networking;
• 46% covered by France 2030;
• open to SMEs and mid-sized companies with revenue above €8m, more than 50 staff and at least 3 years of existence;
• two tracks: Accélérateur IA and Accélérateur IA & Industrie.

These programmes do not replace an internal roadmap. They speed up scoping once the company head has already decided to invest management time.

Three concrete moves for management

1. Choose a narrow scope. One process, one team, one KPI. Examples: handling customer requests, help with sales writing, backlog prioritisation, document quality control. The classic mistake is to “roll out AI” without a named use case.

2. Separate the tool from governance. Catalogue, ambassador, diagnostic: all useful. But the decision stays internal: who validates AI outputs, what data goes into prompts, what level of human oversight, how the gain is measured (time, quality, margin).

3. Keep the European regulatory calendar in view. Since 2 August 2026, the transparency obligations of Article 50 of the EU AI regulation (AI Act) apply to certain systems (in particular, informing users when they interact with an AI). This does not hold the project back: it is a design constraint to build in from the first pilot (notices, welcome messages, traceability of synthetic content where relevant).

What AZINOVE SAS takes from it for its European clients

At AZINOVE SAS, we work with management teams that want useful AI: software development, integration, consulting — not a stack for show. The right order is still:

1. the use case (value and risk);
2. data and processes (where it really takes hold);
3. the tool (sovereign catalogue, European SaaS or custom build);
4. measurement (before / after, no vanity metrics).

If you are an eligible French SME or mid-sized company, the Diagnostic Data IA is often the best funded first step. If you are already past the diagnostic, the challenge is industrialisation: a pilot that survives the consultant’s departure.

Sources (in French)

1. Ministère de l’Économie — Invitation presse « 1 an du plan Osez l’IA », 2 septembre 2026 : https://presse.economie.gouv.fr/ip-1-an-du-plan-osez-lia-jeudi-3-septembre-2026/
2. Direction générale des Entreprises — Le plan national Osez l’IA : https://www.entreprises.gouv.fr/priorites-et-actions/transition-numerique/accompagner-les-entreprises-dans-leur-transition/le-plan
3. Commission européenne — FAQ obligations de transparence article 50 AI Act : https://digital-strategy.ec.europa.eu/en/faqs/transparency-obligations-under-article-50-ai-act