Alice Guéhennec, chief tech, data, and digital officer at Sodexo, oversees an annual technology budget of roughly 500 million euros. A portion of that spending is going toward artificial intelligence tools that automate menu creation, staffing decisions, and pricing-freeing thousands of chefs and facility managers from administrative work that used to eat up weeks of their time.
Menu AI cuts recipe development from weeks to a single day
Guéhennec led the development of an internal tool called "Menu AI" that handles the decisions chefs need to make when creating new dishes. It factors in raw material price swings, seasonality, and demand patterns at each site. The system generates and manages menus from a pool of 400 seasonal recipes in one day, a task that previously required two to four weeks of manual effort. "It's deployed to thousands of sites," Guéhennec said. "All these teams were losing time behind laptops, when their core business is to cook."
Sodexo is now working on adding the ability for AI to scan external food trends from social media and other channels to suggest new recipes, though that capability remains under development. The approach reflects a direct use of AI for Product Development, where speeding up ideation and reacting to market signals are concrete business levers.
Operations get AI-driven staffing and cleaning decisions
On the facilities management side, which accounts for 40% of Sodexo's revenue, Guéhennec has invested in tools that use AI for Operations. These systems help site managers plan staffing levels and track when a room actually needs cleaning, replacing blanket schedules with usage-based triggers. AI also makes data-driven recommendations-telling one location to restock coffee at 4 p.m. to meet demand, while another site can skip the afternoon refill entirely.
Every major AI use case starts with pilots at five to ten sites across a couple of the 43 countries Sodexo serves. Guéhennec said localized feedback during testing "helps us to encourage and facilitate adoption." Employees can also request licenses for AI services like Microsoft Copilot or Google Gemini Enterprise if they see a productivity payoff. "You have to give people freedom with AI if you want them to adopt it," she said. "If it's just a top-down approach, I think it doesn't work."
Frictionless stores and robots test the next frontier
Sodexo is running early experiments with checkout-free retail at two U.S. universities, using AI and computer vision to remove dedicated cashier stations. At one school, the format boosted revenue by 56%; the other saw a 28% sales lift and an 11% rise in average check size. The company has also deployed about 200 robots globally-100 floor cleaners and 85 delivery units, mostly in the U.S.-and is researching kitchen robots that could take on repetitive or risky tasks. "We are looking for some robots like this," Guéhennec said. "To reduce the danger and level of accidents in our kitchens and to replace the less safe tasks with robots."
Why this matters for product development
Product developers across industries wrestle with the same tension: how to balance creative work with the tedious analysis that supports it. Guéhennec's teams cut menu ideation from weeks to a day by handing AI the variables around cost, seasonality, and demand-giving chefs more time for actual cooking. For product professionals, the lesson isn't about the technology itself; it's about identifying the specific, repetitive inputs that slow a team down and offloading them so people can work on the parts of development that need human judgment and craft.
Your membership also unlocks:
