Beyond large groups, unicorns and technological giants, the French economic reality – SMEs – is moving at too slow a speed on AI, a determining subject for their future.
Talking about artificial intelligence in French companies often remains a biased exercise. We look at the large groups, the unicorns, the technological giants. Meanwhile, the economic reality of the country – SMEs – is moving at too slow a speed on a subject which will nevertheless determine their future.
The observation is now well documented. AI adoption in small businesses is two to three times lower than in large ones. In France, only one in three SMEs has initiated uses, often limited to standard tools, without in-depth transformation. And less than 30% of companies declare structured and sustainable use.
This discrepancy is not anecdotal. It marks a future divide in our economic fabric.
Because AI is not an innovation like any other. It acts on all performance levers: productivity, quality, deadlines, customer relations, innovation. Companies that truly integrate it see significant, sometimes spectacular, gains. On the scale of large economies, its impact could exceed one point of additional productivity per year.
But in SMEs, AI is still largely confined to peripheral uses such as the automation of simple tasks, content generation or one-off assistance. It improves the existing without yet transforming the models. However, it is precisely this transformation that will make the difference.
The problem is not access to technology, it has never been easier to test AI tools. The problem lies elsewhere: in the ability to structure an approach, to invest in skills, to rethink the organization. The real subject is the capacity to transform.
The lack of skills, uncertainty about return on investment, costs and technological complexity are all known obstacles. But behind these obstacles lies a more direct reality. AI requires a transformation that many companies have not yet undertaken. Without structured data, without digital culture, without clear vision, AI produces nothing. It amplifies what already exists.
Today, more than one in two managers do not have an AI strategy. Not for lack of conviction, but because the subject goes far beyond the technological question. AI requires us to review processes, professions, sometimes even the economic model.
In SMEs, this transformation almost always rests on the manager. In 73% of cases, it is he who raises the subject. He must understand, decide, and involve his teams, while continuing to run the company in an already complicated macroeconomic context. This reality largely explains the gap between awareness and taking action.
However, time is against us. The risk is not AI per se. The risk is the competitor in France or abroad who uses it better. In many sectors, it already makes it possible to produce faster, at lower cost, with better quality of service. It lowers entry barriers and accelerates the emergence of new players.
For an SME, not getting involved means accepting a gradual dropout – a dropout that is often invisible at the start, but difficult to recover from later.
In this context, the question is no longer whether to adopt AI, but how quickly and with what ambition. This requires moving away from a logic of isolated experimentation to enter into a structured approach: training teams, identifying high-impact use cases, integrating AI into the company’s key functions.
The economic story is clear. Major technological breakthroughs do not benefit those who understand them best, but those who deploy them the fastest.
AI will be no exception.
In three to five years, there will no longer be a question of experimenting or hesitating. It will be a question of staying in the race.
And, for many SMEs, not to get out of it.
