Not a chatbot bolted onto a PDF. An AI menu takes what the guest tells it โ group size, preferences, dietary needs, occasion โ and builds a specific answer: named combinations with dishes, quantities and a total, ready to add in one tap.
For a returning guest it goes further, drawing on what that guest actually ordered before rather than what is generally popular.
menuGPT works from your dishes, your descriptions, your prices and your order history. It recommends what you actually sell and what actually makes you money, not a plausible-sounding dish you took off the menu last year.
Recommendations arrive in the guest's own language across 25+ languages, including full right-to-left layout for Arabic. In UK dining rooms that regularly means Polish, Romanian, Spanish and French in the same service.
Move the sliders to your own numbers. The point is not our figures โ it is how little the uplift has to be before the subscription pays for itself.
5% is the conservative default. WowMenu pilot restaurants have reported considerably more, but your own number is the one that matters.
An estimate based on the figures you entered, not a forecast or a guarantee. Actual results depend on your menu, your guests and how the platform is set up.
A digital menu that actively recommends and personalises rather than just listing dishes โ reading group size, preferences and dietary needs to suggest specific combinations.
No. It recommends on contextual fit and what the guest has signalled, because recommendations that feel appropriate convert and pushy ones do not.
Yes, where a guest profile exists it draws on their previous orders alongside what they say at the table.