France’s small businesses are adopting AI faster than they can measure it
Four in ten small firms report using at least one AI tool, but separate surveys show why adoption, intensive use and proven productivity gains are different claims.
Artificial intelligence has moved quickly into France’s smallest companies. The 2026 France Num barometer says 40 per cent of very small, small and medium-sized businesses use at least one AI solution, up 14 percentage points in a year. Among users, 68 per cent perceive a positive impact on the business.
The survey covers 9,655 companies, including 6,786 very small businesses and 2,869 SMEs, questioned from 23 March to 18 April. That large sample makes the change meaningful, but the measure is self-reported adoption. A company experimenting with a general-purpose writing assistant and one automating a core workflow can both answer yes.
General-purpose tools dominate the first wave
The detail confirms that most use remains accessible and general. France Num says 34 per cent use tools that generate text, voice or images, while 24 per cent use chatbots, assistants or information search. Thirteen per cent report document analysis or classification, 11 per cent task automation and 9 per cent data analysis, forecasting or resource optimisation.
Sector differences are wide. Reported adoption reaches 72 per cent in digital businesses and 60 per cent in specialised services, but 30 per cent in construction, 29 per cent in transport and logistics, 27 per cent in food processing and 18 per cent in agriculture. The barometer also says only 19 per cent of all respondents use paid AI products.
Adoption rates and productivity gains answer different questions
Other datasets explain why the headline should not be read as a productivity result. The Banque de France found that more than two thirds of firms with at least 20 employees used generative AI at the start of 2026, mostly in support functions, while measurable productivity gains remained limited. Its survey covers a different population, so its adoption rate is not directly comparable with France Num’s.
A June Bpifrance survey reported a similar gap among mid-sized companies. Reuters said 77 per cent of 534 company leaders used generative AI, yet only 17 per cent of users reported time savings. Differences in company size, question wording and survey date prevent a single league table, but the studies agree that experimentation is advancing faster than verified return.
The practical test for a small business is therefore not whether it has tried AI. It is whether a defined task becomes faster, more accurate or more valuable after accounting for subscriptions, training, checking and data protection. France’s next barometer will be more useful if it can track intensity, costs and outcomes alongside access. Adoption is now a mainstream indicator; productive integration is still the harder story.
Sources & context
Reporting and reference material used for this article. Context sources do not independently confirm every news claim.
- France Num ↗Primary survey, September 2026: sample, adoption rate, use cases, paid products and sector breakdown.
- Directorate-General for Enterprise ↗Primary government release, updated September 2026: barometer scope and headline acceleration.
- Banque de France ↗Primary research, September-October 2026: generative-AI diffusion and limited measured productivity gains in larger firms.
- Reuters ↗Independent reporting, 23 June: Bpifrance survey of mid-sized firms and reported time-saving results.
Written for WHIF from the linked material. This article does not claim on-the-ground reporting. Our editorial standards.