Winnow launched Winnow Foresight on 29 September 2026, a mobile-first production forecasting web app that predicts kitchen demand at the level of individual dishes. The London-based food waste company built the tool with Hilton chefs, and Hilton is running it while it weighs a wider rollout.
The product is a web app. It opens on any mobile device, needs no extra hardware, and produces a daily production plan from guest occupancy numbers and the site's own history of what was served and what was left. Winnow sells it standalone, or as an add-on for kitchens already running Winnow Vision, the company's waste measurement system.
What Foresight does during a service
Foresight tells the chef how much of each item to prepare before service starts. The plan builds from occupancy figures and past service data, and any number in it can be adjusted before the plan is printed or shared, so anyone on shift can work from it.
After service, the chef records leftovers with a function Winnow calls Snap and Go. A photo of the tray is enough, because Winnow's model estimates the weight without manual scale readings. That result feeds back into the model, so the next forecast uses the previous service's numbers.
Winnow puts the daily routine at 10 to 15 minutes: about five minutes to check the plan, and about ten to record leftovers.
The pilot figures Winnow published
The release names three Hilton properties and their results. All three are Winnow-and-property attributed, and no independent audit accompanies them.
Hampton by Hilton Budapest cut the weight of its leftovers by 69% within six months. The same case carries a 64% reduction in leftover croissants alone, a 4% lift in breakfast satisfaction and a 10% cut in cost per cover.
Hilton London Croydon cut leftover weight by 52% within five months, including a 57% reduction in leftover sausage and a 48% reduction in leftover scrambled eggs.
Hilton Frankfurt Airport recorded a 66% total reduction in overproduction weight and a 5% saving on food costs.
Winnow describes Foresight as currently live in approximately 13 Hilton hotels across Europe, the Middle East and Africa, with Hilton evaluating the results ahead of a potential wider rollout. The 69% figure is the highest in the published set.
The cost case for a hotel kitchen
Winnow's own data puts up to 20% of everything a kitchen buys in the bin. Overproduction drives food cost, labour and disposal spend together. It also carries a climate cost: Winnow reports that its platform customers avoid 122,000 tonnes of CO2 equivalent every year by wasting less.
Marc Zornes, chief executive of Winnow, framed the change as moving from measurement to prevention. "For years, Winnow has helped kitchens measure the food they waste. Foresight goes a step further and helps them see it coming," he said.
Chef Sebastian Nohse, Hilton's senior director of culinary and food and beverage development for EMEA, said the tool takes the guesswork out of production planning for leaner kitchen teams.
Who the tool is built for
Foresight targets focused-service hotels and properties with lighter food and beverage operations. Winnow notes that this segment has had less access to advanced food waste technology than full-service resorts, and that its kitchens often run with smaller back-of-house teams and junior staff on service.
Hilton carries more than 5,000 focused-service hotels worldwide under Hampton by Hilton, Tru by Hilton and Hilton Garden Inn, which is why the group is the reference for the segment. Winnow and Hilton have partnered since 2017, and Winnow Vision now sits in more than 220 Hilton hotels, a footprint Winnow credits with more than 5,900 tonnes of food waste prevented and more than 7.8 million meals kept out of the bin.
What a multi-property group should measure
Foresight includes enterprise-wide reporting, so a group can see performance and adoption across every hotel from one place. The metrics the release ties to the tool are leftover weight by item, cost per cover, breakfast satisfaction and food cost saving.
Where a day-use operation sits in this
A forecast keyed to occupancy numbers and past service data assumes a cover count the property already knows. A day-use block moves that count without adding an overnight guest. A pool pass sold with lunch, a day room taken as a workspace and a daytime meeting block each put heads into a service the overnight occupancy figure does not carry.
For an operator selling hourly and daytime inventory, the practical read is the input rather than the platform. Foresight's plan is only as good as the service history and the occupancy figure fed into it, so a property that runs day-use blocks through a separate system has to decide whether those covers reach the model. The release publishes no day rate, no hourly rate and no minimum stay for any property running Foresight, and it reports no per-block result, so the day-use case has to be tested on the property's own numbers.