Find stuck delivery orders before the customer calls
How do I find orders stuck between the delivery apps and my POS?
The problem, in your words
Orders come in from the delivery apps. Most of them reach the kitchen screen. Some do not. Nobody sees the stuck ones until the customer calls, or the app cancels the order.
The menu goes wrong too. The kitchen runs out of a dish, but the app still sells it. Then someone has to refund it by hand.
What the operator does
The operator is a program that watches your orders all day.
- It checks each order from the apps against what reached your POS.
- It finds orders that are stuck and menus that are out of sync.
- It fixes what your written rulebook allows, and nothing more.
- It asks a person for the rest, with the evidence attached.
- It logs every action with the evidence and an undo.
How we know it works
It grades itself against the right answers, per detector, per day. We tested it on a week of orders it was never tuned on.
| What we measured | Result |
|---|---|
| Precision: of the problems it flagged, how many were real | 0.98 |
| Recall: of the real problems, how many it found | 0.99 |
| Issues it handled alone, without a person | 95% |
The weakest detector is named on the accuracy page.
Three versions you can click
The Gulf version runs on dirhams and riyals with the apps and tills used there. The India version runs the same rulebook on rupees, Mumbai and Bengaluru, Zomato and Swiggy, Petpooja and Posist. The Europe version runs it on euros and pounds, Brussels, London and Madrid, Deliveroo and Glovo, Lightspeed and Toast. Same code, three worlds.
The data is made up
The demo runs on synthetic data. That means we made up the orders to look like a real week. They do not come from a real restaurant. The demo labels the data as synthetic.
On your data
We build the same kind of operator on your apps, your POS and your rules. On day 1 you get a fixed price and 3 to 5 tests. It has to pass them on your real orders.
Contact