AI as EAS: ScanWatch Introduces AI-Powered Article Protection

KAUNAS, Lithuania - Oct. 5, 2023 - PRLog -- Instead of focusing solely on the checkout area, the software can identify what items a particular shopper has picked up from the store shelf.

The new software is currently being tested in a semi-autonomous store setting. In the initial trial the software had managed to cut the theft rate by 42.02%.

Utilizing only the store's existing video security cameras, the program tracks the shopper's movements through the store, noting the items the shopper takes from the shelves and places into the shopping cart or basket. The program then compares this information to what  the shopper actually scans at the checkout to see if all the items taken from the shelves have been paid for.

"For instance, a shopper might pick up three packs of cheese and a bottle of wine but might only scan and pay for the cheese at the checkout. The program would automatically and in real time notify a store employee about the discrepancy or an alarm would sound before the shopper left the store. The program can also be integrated with security gates, which would stay shut if not all the items have been paid for," explained the CEO of ScanWatch, Saulius Kaukenas.

The product is fully compliant with GDPR requirements: no biometric or other personal data is used for shopper identification, and nothing is stored after the customer checks out.

In the words of Mr. Kaukenas, this solution aims to improve the traditional electronic article surveillance framework, where items are equipped with security tags or presented on locked shelves.

"Traditional EAS frameworks require significant hardware and human investments. In recent years, we have seen security tags installed on more and more product categories, including common foods or household goods, requiring more man-hours and increased costs to administer the tagging. This does not scale: hardware-based product security is no longer the answer," said Mr. Kaukenas.

ScanWatch product offerings currently include AI-powered apps to prevent checkout fraud and streamline the self-checkout user experience. Crime Predictor visually inspects every item scanned at the self-checkout and compares the scanned image with the product database. It detects barcode switching, no-scan, no-pay scenarios, as well as mis-scanning of unpacked items (e.g. selecting cucumbers instead of avocados).

Another company product, Picklist Assistant, automatically identifies scanned unpacked products such as fruit, vegetables, or baked goods, thus eliminating the need to manually select  the product from a menu.

Learn more about ScanWatch at https://www.scanwatch.ai/