Tech, ExplainedE-commerce, travel and logistics
Product

Cashierless store

Take it and leave; you are charged.

BEFORE ITAFTER ITCashierless storeCOVERSCeiling camerasShelf sensorsAutomatic chargingNo counterVideo analytics↗ SMART BUILDINGStrackseach shopperEdge AI↗ GENERATIVE AI APPLICATIONSrecognisestaken itemschargesthe shopperPayment gateway↗ ONLINE PAYMENTSrecordsthe saleRetail POSCONCEPTS IT USESComputer visionBEFORE ITAFTER ITCashierless storeCOVERSCeilingcamerasShelf sensorsAutomaticchargingNo counterVideo analytics↗ SMART BUILDINGStrackseach shopperEdge AI↗ GENERATIVEAI APPLICATIONSrecognisestaken itemschargesthe shopperPayment gateway↗ ONLINE PAYMENTSrecordsthe saleRetail POSCONCEPTS IT USESComputer visionBEFORE ITVideo analytics↗ SMART BUILDINGS↓ tracks each shopperEdge AI↗ GENERATIVE AI APPLICATIONS↓ recognises taken itemsCashierless storeCeiling camerasShelf sensorsAutomatic chargingNo counterAFTER IT↓ charges the shopperPayment gateway↗ ONLINE PAYMENTS↓ records the saleRetail POSCONCEPTS IT USESComputer vision
The cashierless store follows a Gurugram employee from the gate, records the buttermilk and sandwich she takes from the shelves, and charges her ₹100 after she walks out.
opens a page inside this one a page in the same scene a page in another scene a concept page a label only, with no page

A cashierless store is a shop with no billing counter at all. Cameras and shelf sensors record what each shopper takes, and the shopper is charged automatically after walking out. It is also called a checkout-free store, or a just-walk-out store. Most such stores so far are small convenience shops, in offices, airports, stadiums and stations.

Here is how it works. The shopper first identifies herself at the entry gate, usually by scanning a QR code in the store's app. The app is linked to a payment method. Cameras on the ceiling then follow the shopper through the store. They do not need to recognise her face. They only need to follow the same person from the gate to the exit. Each shelf also has weight sensors. When an item is picked up, the shelf's weight falls by that item's weight. When the item is put back, the weight rises again.

The store's software combines the two. The cameras show who stood at the shelf, and the shelf shows which item left it. This uses computer vision, software that recognises objects and people in images, together with the sensor data. The cameras produce a very large amount of video. So most of the processing is done by computers inside the store, close to the cameras, rather than in a distant data centre. This is called edge AI. Similar video analytics software is used in other places, for example to count people in a building.

Imagine a small cashierless shop on an office campus in Gurugram. An employee scans the app at the gate at 4 pm. She takes a ₹40 bottle of buttermilk and a ₹60 sandwich. She picks up a ₹30 chocolate but puts it back. Then she walks out. Two minutes later, the app shows her a bill for ₹100. The store charges her saved payment method through a payment gateway. The sale is also recorded in the store's retail POS back office, like any other sale, so the stock is updated.

The difficult problems are cost and mistakes. The many cameras, shelf sensors and in-store computers cost a lot for a small shop. When several shoppers stand close together, the system can charge the wrong person. So the shopper can dispute an item in the app. Retailers weigh these costs against the cashiers they no longer need. Many stores choose a simpler step instead: a self-checkout kiosk, or scan and go on the shopper's own phone.

Tech, Explained is a part of ingrid.fyi. Every page explains one piece of technology in plain words.