Artificial intelligence has slipped into actions you already perform without thinking. When your email inbox sorts spam, when a navigation app recalculates your route in real-time, when an online store suggests a product: behind each of these actions, an AI algorithm processes data on your behalf.
In businesses, the movement is accelerating. According to Insee (Insee Première n° 2120, published on July 21, 2026), 18% of French companies with 10 or more employees used at least one AI technology in 2025, compared to 6% two years earlier.
What the AI Act concretely changes for French companies
Competitors talk a lot about productivity. They almost always forget the legal framework that now conditions the use of AI in Europe. The AI Act, the European regulation on artificial intelligence, is not a theoretical text: some of its obligations already apply.
Since February 2, 2025, AI practices deemed to pose unacceptable risk are prohibited. This includes, for example, social scoring systems or large-scale behavioral manipulation. Companies deploying or using these tools must ensure a sufficient level of AI mastery within their teams.
Since August 2, 2026, a second wave of obligations has come into effect. It concerns transparency: chatbots must indicate that they are automated systems, and certain AI-generated content (images, videos) must be identified as such. For an SME using a conversational assistant on its website, this implies clear labeling, under penalty of sanctions.
At Botagora, this interplay between technology and compliance is part of the topics addressed to support companies in their AI-related projects.
Generative AI at work: real gains but very uneven across professions
Have you ever asked a tool like Copilot to draft a meeting summary or rephrase an email? This type of use of generative AI is spreading quickly. The tools exist, they are accessible, and many employees adopt them without waiting for directives from their management.
The time savings are real for certain tasks: writing reports, summarizing long documents, generating initial drafts of computer code. However, the benefits vary greatly from one profession to another. A developer using a coding assistant can speed up certain phases of work. A salesperson relying on AI to personalize their follow-ups gains in responsiveness.
For other functions, the gain remains marginal or even nonexistent. Manual jobs, positions with a strong relational component, or activities requiring fine contextual judgment do not derive the same profit from these tools. AI does not replace field expertise. It accelerates information processing when the information is structured and abundant.
What distinguishes successful adoption from a passing trend
Deploying a generative AI tool in a company is not enough. Without training or a framework for use, adoption remains superficial. Companies that derive measurable benefits from AI are those that have defined specific use cases, trained their teams, and established processes for verifying the results produced by the machine.
A concrete example: entrusting the drafting of a contract to a generative tool without legal review exposes the company to factual errors or inappropriate clauses. AI proposes, the professional validates. This division of roles is the condition for the tool to serve rather than harm.
AI adoption in France: the figures show a persistent lag
The progress is clear: from 6% in 2023 to 18% in 2025, the number of French companies using AI has tripled. This figure, from Insee, deserves to be viewed with perspective. France remains below the European average, estimated at 20%.
This gap is partly explained by the structure of the French economic fabric, which is mostly composed of SMEs. Large companies and mid-sized enterprises adopt AI faster because they have dedicated budgets, large internal data, and technical skills in-house. For a 15-employee SME, the main barrier is not the cost of the tool: it is the time required to identify a relevant use case and train the team.
What the government plan provides for SMEs
Several public initiatives aim to support SMEs in integrating artificial intelligence, with service-based approaches (accounting, customer relations, logistics) rather than a global approach often too abstract for small structures.
The stated goal is to make AI accessible without requiring advanced technical skills. It remains to be seen whether these initiatives will effectively reach the companies that need them most, those without a technical director or innovation budget.
Concrete uses of artificial intelligence in daily life
Beyond the business world, AI is changing domestic and personal habits. Here are the most common uses, often employed without identifying the technology behind them:
- Voice assistants (like smart speakers) that interpret natural language commands to control smart home devices, play music, or answer factual questions.
- Health applications that analyze sleep, heart rate, or physical activity data to produce personalized recommendations.
- Instant translation tools capable of translating spoken conversations in real-time with a level of accuracy that improves each year.
- Recommendation systems of streaming platforms that learn from your choices to refine their suggestions.
All these uses rely on the same principle: an algorithm trained on large amounts of data learns to recognize patterns and produce appropriate responses. The difference from a classic program is that the system improves over time and with the data it receives.
Artificial intelligence is neither a magic wand nor an abstract threat. It is a tool whose usefulness directly depends on how it is integrated, whether in a personal routine or in a company’s processes. Companies that benefit from it are those that train their teams and define specific use cases. Others accumulate subscriptions to tools that no one really uses.
