Winnow Foresight Cut Hilton Food Cost 5% by Forecasting Kitchen Demand

Hampton by Hilton Budapest cut leftover food weight by 69 percent in a pilot of Winnow Foresight, a mobile-first AI forecasting tool launched this week for hotel kitchens. Hilton London Croydon reduced leftovers by 52 percent, and Hilton Frankfurt Airport cut overproduction by 66 percent while saving 5 percent on food costs. The pilots ran across 13 European Hilton properties.

The tool was developed with Hilton’s culinary team and aimed at focused-service hotels, which typically lack access to advanced waste-reduction technology. Foresight analyzes occupancy data and historical service records to produce item-level production forecasts, so the kitchen prepares the right amount from the start rather than trimming waste afterward. After service, chefs photograph the leftovers and the AI estimates their weight, which refines the next forecast. Winnow CEO Marc Zornes described the product as predicting demand at the item level.

For an operator, the number to read is the 5 percent food-cost saving at Frankfurt Airport, not the headline waste percentages. Food cost sits in the flow-through math of a hotel kitchen, and a focused-service hotel with a thin kitchen team has few other levers on it. Weight reduction is a sustainability metric; the P&L effect is what a general manager can defend to ownership.

Two caveats before anyone extrapolates. These are results Winnow reports from pilots at three named Hilton hotels, not a portfolio-wide audit, and the save depends on the daily discipline of photographing leftovers so the model keeps learning. Ask any vendor in this category what the labor cost of that data capture is per shift.

Winnow also states its existing technology prevents more than 128,000 tonnes of CO2-equivalent emissions a year across over 3,500 kitchens worldwide. The signal for hospitality is that AI adoption is reaching back-of-house operations, where margins are tight and analytics support is scarce.