Baked too much
10:00pm
Trays still full at closing. Good bread goes in the bin.
Prototype for UAE bakeries
خبزك ينباع، والباقي ما ينرمى.
Ghadan is an AI assistant for independent bakeries. Every night it sends tomorrow's baking plan on WhatsApp, item by item, with the reason behind each number. Before closing, it helps you sell what is left.
01The problem
Most bakeries decide tomorrow's quantities the same way: bake what we baked last week. When the day turns out different, the guess is wrong.
Baked too much
10:00pm
Trays still full at closing. Good bread goes in the bin.
Baked too little
2:00pm
Croissants sold out by lunch. Customers leave, and some don't come back.
The UAE's national initiative Ne'ma aims to halve food waste by 2030. Ghadan starts where bakery waste starts: the night before.
02How it works
Ghadan learns how each item sells: by day of week, weather, Ramadan, Eid, holidays and promotions.
You getA model for every item
Tomorrow's quantities on WhatsApp, each with a short reason. Reply ✅ or change a number.
You getA clear plan before baking
If too much will be left, Ghadan suggests a discount until closing. Approve with one tap.
You getA ready post for your customers
Log what was left: donated, binned or sold on offer. Big order or broken oven? Tell Ghadan and the plan adjusts.
You getA waste and donation log
03Try it
Pick any date and Ghadan fills in the occasion, weather, school holidays and payday by itself. Change anything and watch the plan move. This is the real model running in your browser, trained on a sample Sharjah bakery.
Anything unusual?
Tomorrow's plan
04Tested
Sample data · not a real bakery yet
less waste, over one full year of a sample Sharjah bakery, compared with baking the same as last week.
We generated two years of sales for a sample Sharjah bakery, with weekdays, weather, Ramadan, Eid, holidays, promotions and big orders. For every day of the last year, the model trained only on the days before it, then its plan was compared with what customers wanted. The baseline bakes the same amount as on the same day last week. Real bakeries will differ, and that is exactly what the pilot is for.
Perspective
10 croissants
left every night
×
7 AED
each
×
365
nights
25,550 AED
a year at shelf price. From one item.
An example to show the math, not figures from a real bakery.
05The prototype
FAQ
Ghadan is a working prototype. The forecast engine, the dashboard and the WhatsApp flow all run today on a sample bakery. We are looking for the first UAE bakeries to pilot it with.
Every number comes with its reason, like "Hot day, 8% less". You can change any number before baking. Ghadan suggests, you decide.
Yes. It learns how each item changes during Ramadan, the last 10 nights and Eid, and how sales move to the evening. Islamic dates in the prototype are approximate.
The idea is that you don't. Ghadan is designed to run from simple daily numbers, what you baked and what sold, and to talk to you on WhatsApp. Connecting to point of sale systems comes later.
The messages are designed and work inside the prototype as previews. Sending them through the WhatsApp Business platform is the next step.
First, Ghadan tries to bake less of it. Before closing, it suggests an offer. What still remains you log as donated, binned or sold, so the next plan gets sharper.
Join
We are starting with a small pilot for independent bakeries in the UAE. Tell us about yours and we will reach out on WhatsApp.