VivaTech: AI Escapes Companies, Self-Taught Employees

The Enthusiasm of VivaTech vs. Office Reality

At the latest edition of VivaTech, business leaders were captivated by demonstrations of technological innovations, particularly in the field of artificial intelligence (AI). Concepts such as autonomous agents, multimodal models, and intelligent copilots sparked keen interest, hinting at an imminent transformation across various sectors. However, this enthusiasm starkly contrasts with the situation in offices, where the adoption of these technologies remains limited.

A study conducted by Master The Monster-Kantar reveals that 70% of employees using AI have acquired their skills through self-learning, while only 9% have received training provided by their company. Furthermore, 90% of employees believe that their company lacks a clear strategy regarding AI, and 50% feel that their organization is lagging technologically.

The Gap Between Innovations and Business Use

The spectacular advancements in AI do little good for companies if their employees are not trained to use them effectively. Currently, businesses seem hesitant to invest in training, which hinders the adoption of these technologies. As a result, employees are taking the initiative to learn on their own, often using personal tools to accomplish their professional tasks.

In France, 46% of workers use their own AI tools for work, a practice known as Shadow AI. While this poses security risks, the real issue lies in the fact that companies do not provide the necessary resources for optimal AI use. Only 25% of employees believe that the tools provided by their company meet their needs, prompting them to seek faster (23%), simpler (16%), and more effective (14%) alternatives.

The Economic Impact of Untapped AI

This situation is not without consequences for organizations. Although the use of AI allows employees to save time, companies fail to capitalize on these gains. A 2026 study conducted by researchers from Stanford and Carnegie Mellon highlights that generative AI generates a value of $172 billion per year for its users in the United States, while revenues for AI-focused companies only reach $14 billion. Thus, the value created by AI is primarily captured by individuals, not organizations.

When an employee saves two hours a week using tools like ChatGPT, without their employer knowing or organizing it, these extra hours benefit only the individual and, indirectly, the shareholders of OpenAI. This leakage of value, which no one has yet bothered to stop, represents a significant loss for companies.

Transatlantic Comparison: France vs. the United States

The issue of AI training is not unique to France. In the United States, 42% of employees report that their employer expects them to learn AI on their own. However, the major difference lies in the involvement of senior management. In the U.S., 65% of managers use Shadow AI, compared to 31% of employees. This shows that American management is adopting these tools, even in a disorganized manner, while in France, it is primarily employees who are tinkering with AI.

This difference in approach reflects a problem of engagement in France, as opposed to a problem of alignment in the United States. The AI market, dominated by American models and platforms, risks leaving French companies behind if they do not take swift action to integrate these technologies.

The Urgency of Strategic Transformation

This debate goes beyond mere productivity. It touches on the fundamental transformation of the world of work. According to Arthur Mensch, CEO of Mistral AI, engineers no longer code but orchestrate agents to perform tasks. This radical change requires a complete reinvention of organizations.

AI training is no longer just about improving current productivity but also about the survival of French companies in an environment where jobs are rapidly evolving. Transitioning from execution to orchestrating agents represents a major challenge that requires a clear strategy.

Companies must address the questions posed by AI, such as the role of middle management, the integration of young talent, decision-making autonomy, and the redefinition of skills. Time is of the essence, and transformation must be swift to avoid being outpaced by American companies.