Moderated by Pascale Davies (Euronews), the panel discussion featured Vincent Faburel (Interflora), Alexandra Lorenzo de Brionne (Michelin), and Abdallah Arioua (former Chief Data & AI Officer at Relyens | Entrepreneur).

Here are the five key takeaways from the discussions.

1. AI creates value only when it addresses a business need

The most promising projects aren’t necessarily the most innovative ones. Above all, they address a clearly identified business challenge. The speakers noted that initiatives driven solely by technology often struggle to move beyond the experimental stage.

2. Data governance remains the foundation of any AI strategy

Before discussing models, we need to address data quality, definition, and governance. Without reliable, shared, and properly defined data, it is impossible to sustainably scale an AI project.

3. Knowing when to stop a project is also a sign of maturity

Not all use cases are intended for production deployment. Discussions have shown that a mature organization also knows how to recognize when a project does not create enough value and to choose to redirect its efforts.

4. Models evolve quickly, but organizations do so much more slowly

Generative AI is advancing at a spectacular pace. The real challenge now lies in integrating it into business processes while balancing security, compliance, cost control, and user adoption.

5. AI in production is, above all, an execution challenge

The main takeaway from the evening is undoubtedly this: models are now accessible to everyone. What makes the difference today is companies’ ability to turn this technology into tangible value, to support change, and to measure its impact over time.

To learn about the evening’s highlights, hear feedback from our speakers, and get a more in-depth analysis of the discussions, check out our dedicated article on Techerati.

Pre-register for Tech Show Paris 2026 to meet the leaders shaping the future of AI, data, the cloud, and cybersecurity.