How does a €4 billion food group with 11,000 employees rebuild its entire digital foundation in six years? At Tech For Industry Show 2026 (Paris Expo Porte de Versailles), [speaker to be confirmed] presented the transformation of Bel Group — the cheesemaker behind The Laughing Cow, Babybel and Kiri. Behind the usual showcases of R&D, marketing and commerce lies a far larger undertaking: 25 major digital transformations, covering every function and every market.

The starting point: an information system inherited from 650 small individual platforms, "democratically spread" across all functions. Bel is replacing them with one fully interconnected core made up of an ERP, a PLM and an MES. A massive change aimed at finally connecting the entire organization around the latest technology.

A striking admission: until three years ago, no work had been done on data foundations at Bel. In three years, the data team has grown from 5 to 55 people, with data governance, data structures and "data citizenship" now in place — the condition for all this digitization to actually work.

Another finding: 95% of Bel's core processes were not mapped, let alone optimized. They lived in employees' memories — with the risk that a retirement could walk a critical process out the door. Yet, the speaker stresses, "artificial intelligence doesn't work very well without structured processes." So Bel is putting a process architecture in place function by function, backed by an "army" of around 100 Business Process Owners (BPOs).

This is Bel's singularity: most companies accelerate on AI after already having their technology and data layers in place. Bel has to install all three at once — technology, data and AI. More complex, more stressful for teams and partners, but with a benefit: building from the outset an ecosystem able to work seamlessly on the latest technology. All while managing change for 11,000 people and delivering on budget.

The digital roadmap is tied to the strategic plan, with buy-in and co-ownership from all functions at executive committee level. The result: the speaker is challenged on value or budget increases, but never on the "why" — the priority is shared by everyone.

A key message: there is no compromise to be made between sustainability and profitability. "It's and, not or." A solution meeting only one of the two is the wrong solution — back to the drawing board. Running a company on profitable growth alone would be easier; doing both is hard, but possible, provided you look hard enough for the win-win.

Bel relies on Dassault Systèmes (software for its MES and PLM, and collaboration on creating the virtual twin) and on Accenture, which supports all three layers of the transformation — technology, data and AI.

Beyond its ambition, what the speaker wants remembered about Bel's transformation is its purpose: facing a "broken" food system, making a contribution — however large or small — to fixing it, so future generations have a livable future, away from summers of 40–45°C. AI offers unprecedented power; the real question is what we do with it. The closing words: let's agree to do something good.

Why map your processes before deploying AI? Because, as the Bel case illustrates, AI works poorly without structured processes. Unmapped processes live in employees' memories and vanish when they leave, weakening both the organization and any AI project.

Can you accelerate on AI without data foundations? With difficulty. Bel chose to lay technology, data and AI in parallel, but stresses that most companies have the advantage of already having their tech and data layers before they start.

Are sustainability and profitability compatible? Yes, according to the keynote: Bel refuses any compromise between the two and considers that a solution meeting only one criterion must be reworked. It's harder, but a win-win solution exists if you look hard enough.

This conference was recorded at Tech For Industry Show 2026, the B2B trade show for industrial transformation (Smart Manufacturing, AI, Supply Chain) at Paris Expo Porte de Versailles. Watch the full keynote video above, and join us on May 19–20, 2027, Hall 7, for the next edition.