SMEs, Mid-Caps, Regulated Sectors: France's AI Gap Isn't the Same Everywhere
72% of French SME leaders have no clear AI plan. The real gap varies by company size and sector, with very different risks for SMEs and mid-caps.
72% of French SME leaders have no clear AI plan, and only 11% of French SMEs show advanced AI usage (Bpifrance Le Lab, 2026; MIT Sloan Management Review). That headline figure hides a finer reality: the gap looks nothing alike whether you are talking about a 20-employee SME or an 800-person mid-cap operating under regulatory obligations.
In my engagements, I see two distinct patterns of delay. An unregulated SME moves fast, with no safety net: teams adopt an AI assistant within days, without any validation or scoping, feeding exactly the shadow AI I observe on the ground, 40 to 60% of real usage escaping IT oversight entirely. A mid-cap in a regulated sector, finance, healthcare, industry, moves more slowly, with more process in place, but its legal exposure is far greater the moment an undocumented use touches sensitive data or an automated decision-making step.
This difference carries three concrete consequences. First, budget waste: AI licenses purchased separately by different departments with no coordination, often redundant. Second, a growing legal risk: 80% of SMEs have no documented AI governance, a gap that will get costly once the AI Act's stricter obligations for high-risk systems fully apply. Third, a quiet competitive lag: 73% of AI integration attempts fail or get abandoned within the first six months for lack of initial scoping (Gartner, 2024), while better-structured competitors consolidate their lead.
This is exactly what the first stage covers, the usage inventory, in week 1: mapping real AI usage before any tooling decision, SME or mid-cap, regulated or not. It is also the first of the eight governance principles I defend on nachnouchi.com, governance before deployment, never the reverse. An SME and a mid-cap do not need the same level of formality, but neither can afford to scope after the fact what it has already let run unchecked.
France's AI gap will not close with more tools, it closes with an honest diagnostic of what is already in place, whether it works or not, sized to the company's scale and sector. If you'd like to discuss it, get in touch.
Frequently asked questions
What is the difference in AI readiness between a French SME and a mid-cap company?
An SME moves faster but with no safety net, which fuels shadow AI, while a mid-cap moves more slowly but faces much greater legal exposure once it belongs to a regulated sector such as finance or healthcare. This is a field observation from nachnouchi.com engagements, not an external quantified study.
What percentage of SME leaders have no clear AI plan?
72% of French SME leaders have no clear AI plan (Bpifrance Le Lab, 2026).
How many French SMEs have advanced AI usage?
Only 11% of French SMEs show advanced AI usage (MIT Sloan Management Review).
What is shadow AI and why does it mainly affect SMEs?
Shadow AI refers to AI usage not disclosed to leadership or IT. 40 to 60% of real AI usage in SMEs and mid-caps escapes IT oversight, a field observation from nachnouchi.com rather than an external quantified study.
Why are regulated sectors more exposed to the AI gap?
Because undocumented use touching sensitive data or an automated decision-making step directly exposes the company to AI Act and GDPR obligations, while 80% of SMEs have no documented AI governance (nachnouchi.com).
Why do so many AI projects fail in SMEs and mid-caps?
73% of AI integration attempts fail or get abandoned within the first six months, most often for lack of clear initial scoping (Gartner, 2024).
What is the first step to closing this gap?
The Inventory stage, in week 1, which consists of mapping the company's real AI usage before any tooling decision, regardless of its size or sector.
Should an SME and a mid-cap follow the same AI governance approach?
The level of formality differs by size and sector, but the principle stays the same: governance before deployment, never the reverse.