What criteria allow us to distinguish a truly useful IT resource from a tool quickly abandoned? The question arises with new urgency: 26% of French TPE-PMEs report using AI in 2025, which is twice as many as a year earlier, according to France Num. This acceleration reshapes the landscape of IT resources mobilized in digital projects, between widely adopted free tools and still minority paid solutions.
AI Adoption and Free Tools: What Recent Data Reveals
| Indicator | Data | Source |
|---|---|---|
| TPE-PMEs using AI in 2025 | 26% | France Num 2025 |
| Growth compared to 2024 | Doubling in one year | France Num 2025 |
| Share of generative AI users on free tools | About half | Bpifrance Le Lab, January 2026 |
| Shift towards structured paid offers | Minority of users | Bpifrance Le Lab, January 2026 |
The table highlights an imbalance. The pace of adoption is rapid, but half of the TPE-PMEs using generative AI rely on free and generic tools. The use of paid solutions, with technical support and sustainability guarantees, remains limited to a smaller fraction.
This gap has direct consequences on the robustness of digital projects. A free tool can disappear, change its terms of use, or limit its functionalities overnight. Building a workflow around a resource whose continuity is not assured weakens the entire project.
To assess the reliability of a resource before integrating it into a project, D’clic Lab’s IT resources allow for comparing different categories of solutions based on their professional use.
European AI Regulation: Regulatory Constraints to Anticipate Starting in 2026
The AI Act, the European regulation on artificial intelligence, imposes new obligations on companies starting August 2, 2026. This deadline directly concerns the choice of IT resources integrated into a digital project.
AI systems classified as high-risk (automated recruitment, credit scoring, infrastructure management) will need to comply with requirements for transparency, traceability, and human oversight. AI-generated content must be clearly labeled as such.
- Generative AI tools used to produce text, images, or code must clearly indicate the artificial nature of the produced content.
- High-risk classified systems will require detailed technical documentation and a human oversight mechanism.
- AI practices deemed to pose unacceptable risks (behavioral manipulation, social scoring) are prohibited.
For a TPE-PME choosing its IT resources, verifying a tool’s compliance with the AI Act becomes a selection criterion. A project management software that integrates AI to prioritize tasks does not raise the same regulatory questions as an automated application analysis tool.
European SaaS solution providers are beginning to communicate about their compliance. In contrast, free tools from outside Europe rarely provide this information, which increases the legal risk for the user company.
IT Resources for Digital Projects: Three Often Overlooked Selection Criteria
Lists of digital tools available online focus on functionalities. However, three less visible criteria determine the long-term viability of an IT resource in a project.
Hosting and Data Location
A tool hosted in Europe offers a more predictable legal framework for data protection, especially since the strengthening of the GDPR. The location of the server determines the level of control over the project’s data. Some collaborative suites like MyUniSpace emphasize European hosting without tracking or reselling data.
Interoperability with the Existing Ecosystem
A high-performing project management software that is isolated from the rest of the chain (messaging, storage, billing) generates duplicates and wastes time. Interoperability through open APIs or native connectors reduces friction.
Before adopting a new tool, checking if it integrates with already used applications avoids manually recreating bridges. This technical point, rarely highlighted in comparisons, makes the difference between a tool that is sustainably adopted and one that is abandoned after a few weeks.
Business Model and Sustainability
A freemium tool whose model relies on monetizing user data does not offer the same guarantees as open-source software supported by an active community. Penpot (design and prototyping) or Goose (AI agent for code) illustrate documented and maintained open-source alternatives.
Cybersecurity of Digital Tools: A Blind Spot in Digital Projects for SMEs
The acceleration of digital tool adoption in SMEs does not always come with an increase in cybersecurity skills. Multiplying applications, especially free ones, broadens the attack surface.
- Each new tool connected to the company’s information system represents a potential entry point for intrusion.
- Reusing passwords between free tools and professional accounts remains a common vulnerability.
- The absence of end-to-end encryption on certain messaging or storage solutions exposes sensitive project data.
Auditing the security policy of a tool before integrating it into a project takes just a few minutes. Checking for the existence of two-factor authentication, data encryption at rest and in transit, and a notification policy in case of a breach is a minimum.
Companies that have structured their digital transition around solutions evaluated on these criteria experience less friction in use and fewer incidents. The choice of an IT resource for a digital project is not limited to its displayed functionalities: regulatory compliance, data location, and the robustness of the business model weigh as heavily as usability in medium-term success.