The right dish, while they are still deciding.
Suggestions in the menu and at the cart, drawn from what this guest has ordered before and what tables like theirs order together — with dietary filters treated as absolute.
- In-session, not after
- Dietary filters are hard
- Tracked to the order
What it does
Turn what you learn into more covers.
- While the decision is open
- The suggestion arrives on the menu and in the cart, when the guest is still choosing — not in an email the following week.
- Dietary rules are not preferences
- A filtered-out dish is never recommended, ever. That rule sits above every other signal in the engine.
- It works on a first visit
- New guests get suggestions built from what this restaurant’s tables actually order together, so there is no cold start where the menu has nothing to say.
- Counted honestly
- A recommendation is credited only when it was actually shown and then taken. Shown, clicked and added are three separate numbers.
See it
Example data
- Shown100%2,140
A suggestion appeared on a menu or in a cart
- Opened23%486
The guest tapped through to the dish
- Added8%173
It went into the order
Three separate numbers, not one. A recommendation that was never actually on screen cannot take credit for the order that followed it — which is what makes the last figure worth anything.
Example fortnight · dietary filters applied before anything is suggested
Where it sits
One system, three jobs
Dish recommendations is part of grow.
Turn what you learn into more covers. It runs on the same menu, the same tables and the same bill as everything else here — nothing to integrate, nothing to reconcile.
Take the order without anyone at the table.
QR ordering · Digital menus · Live orders · Kitchen display · Guest payments
Know where every pound came from.
Revenue & attribution · Analytics · AI insights · Guest intelligence · Automated messaging