Task
Cut queues in the headquarters canteen at peak hours and speed up checkout without cashiers.
The business-lunch menu changes every day, so the system has to learn new dishes quickly.
Sber
~98%
AI checkouts at Sber headquarters
Recognition accuracy.
Computer visionEdge inferenceSelf-checkoutsFoodscanner
~98%
recognition accuracy
300+
receipts per checkout per business lunch
~10 min
learning a menu of 150+ items
Four Foodscanner-based checkouts in the canteen of the office on Kutuzovsky : a camera recognizes the dishes on a tray and issues the receipt right away.
Client
Sber
Industry
Foodtech
Areas
Computer vision, self-checkouts
ESTESIS role
Deployment of the Foodscanner product
From task to result
Client
Sber is the largest bank in Russia. The checkouts are installed in the canteen of its headquarters on Kutuzovsky Avenue.
What we did and how
We deployed an ESTESIS product — the Foodscanner AI self-checkout.
- 01We installed 4 self-checkouts
- 02A camera above the tray recognizes the dishes and issues the receipt
- 03Learning a new menu of 150+ items takes about 10 minutes before the shift
- 04Working through the business-lunch rush without a cashier
Result
~98%
recognition accuracy
300+
receipts per checkout per business lunch
~10 min
learning a menu of 150+ items
Each checkout rings up more than 300 receipts per business-lunch shift with a recognition accuracy of about 98%.
Stack and competencies
More stories
Your task —next story
Tell us about the project — within 5 working days we will send an architecture and an estimate.
Accredited IT company · Russian Ministry of Digital Development registry No. 20142