An industrial SME in the Grand Est has deployed an artificial intelligence tool to sort its quote requests. The result: processing time has dropped from several days to just a few hours, without any additional hiring. This type of digital transformation does not rely on a colossal budget or a complete overhaul of the business. It starts from a specific operational irritant, and that’s often where everything is at stake.
Regulatory obligations related to AI: what the AI Act changes concretely
Since August 2, 2026, the European regulation 2024/1689 (AI Act) imposes transparency obligations on certain AI systems, particularly those that generate content or interact directly with users. In practice, if your company uses a chatbot or a text generator in customer relations, you must inform the user that they are interacting with an AI.
The timeline for obligations related to high-risk AI systems has evolved, making continuous regulatory monitoring necessary. We are not talking about a distant text: the first sanctions already concern common uses. Ignoring this framework exposes you to fines, but also to a loss of customer trust that is difficult to regain.
For leaders of SMEs and mid-sized enterprises, the reflex to adopt is to map out the AI tools already used in the company (including those adopted by employees without formal validation) and to check their compliance. You can find more information on make-world.org to structure this approach and identify resources suitable for small organizations.
Adoption of AI in French SMEs: optimize before innovating
According to the France Num 2025 barometer, conducted by CREDOC for DGE, very small and medium-sized enterprises are adopting artificial intelligence more quickly than expected. The TIC 2025 survey by Insee confirms the trend: the information and communication sector reaches 59% AI usage.
The on-the-ground observation is more nuanced. The majority of adopting companies use AI to improve existing processes (automating repetitive tasks, sorting data, drafting reports) rather than to create new business models. This is not a failure; it is a logical starting point.
Three concrete uses that transform daily operations
- Automation of email processing and incoming requests: an AI filter redirects each message to the right contact, reducing response times and freeing up sales time
- Predictive stock analysis based on sales history: avoiding overstocking without risking shortages, a direct lever on cash flow
- Assisted generation of marketing content (product sheets, descriptions, social media posts): the time savings are significant, provided each output is proofread and adapted
The trap would be to multiply tools without a clear process. A single tool well integrated into an existing workflow produces more value than five underutilized subscriptions.
Cybersecurity and digital transformation: the link that SMEs neglect
The digitization of processes mechanically expands the attack surface. Cloud tools, collaborative platforms, and connected AI solutions create numerous potential entry points for intrusion. Competitors often talk about strategy and digital culture, but rarely about what happens when a ransomware attack halts operations for a week.
The NIS2 directive strengthens cybersecurity requirements for a broader range of businesses in Europe. Failing to integrate security from the outset of a digital project is akin to building without foundations.
Priority actions on a limited budget
You don’t need a SOC (Security Operations Center) to start. Three actions already significantly reduce risk:
- Activate multi-factor authentication on all professional accounts (email, CRM, cloud tools): it’s free on most platforms and blocks the majority of phishing attempts
- Train teams once a quarter on basic reflexes (not clicking on suspicious links, verifying the sender, reporting any anomalies): feedback varies on the effectiveness of formats, but a phishing simulation exercise remains the most impactful
- Back up critical data on a network-disconnected medium, tested regularly: in the event of an attack, the ability to restore determines whether the company resumes operations in a few hours or several weeks
Change management: starting from the ground rather than from PowerPoint
Bpifrance Le Lab already pointed out that the main barrier to digital transformation for SMEs and mid-sized enterprises is the perceived complexity of the topic. The problem is not technical; it is human. A digital project that does not solve a problem felt by the teams will be circumvented or abandoned.
The method that works on the ground involves identifying a specific operational irritant in a service (a manual reporting process that takes two days, an email validation process that generates errors) and addressing it with a digital tool while directly involving the concerned users.
This first visible success creates a snowball effect. Employees see the concrete gain, share their experience, and resistance to change naturally decreases. Then you move on to the second irritant, and then the third. Digital transformation progresses through accumulated evidence, not by decree.
The connected customer, more demanding and more volatile, also drives this transformation from the outside. The digital customer experience (response times, personalization, smooth purchasing journey) becomes a selection criterion on par with price. A company that digitizes its internal processes without rethinking its customer journey misses half the point.
Transformation in the digital age is not a project with an end date. It is a continuous capacity for adaptation, built brick by brick, from real problems and measurable results.
