Swiss smart data capture company Scandit has announced real-time vision AI to help retailers detect loss patterns, whether accidental or intentional, at self-checkout as they happen.

It empowers shoppers to self-correct, without store staff intervention in most cases, to keep the checkout flow uninterrupted.

Scandit’s entry into the loss prevention market addresses one of retail’s most visible tensions: reducing loss at self checkout without adding friction that frustrates shoppers or burdens store teams. Self-checkout now handles 54% of transactions in stores that have it. Those stores carry losses 33% higher than comparable stores without it, and loss rises an average of 22% in the year after installation. Random spot checks can take in excess of two minutes causing customer friction and absorbing staff time.

With Scandit’s self-checkout loss prevention, a soft on-screen nudge to shoppers, or attendant notification if escalated, corrects any flagged items before a transaction is completed, resulting in the recovery or deterrence of more than 75% of losses attributed to self-checkout. This includes known loss patterns such as missed scans, intentionally skipped items, left in trolley and abandoned transactions.

“Self-checkout fundamentally changed retail, but it also changed the loss equation,” said Christian Floerkemeier, Scandit CTO. “Retailers shouldn’t have to choose between a fast checkout experience and strong loss prevention. Self-checkout loss prevention gives them a software-first way to detect issues in real time with the process configurable to a retailer’s own policy. In addition to recovering losses, it also enables retailers to optimise labour usage more productively.”

The software is deployed as a per-station architecture that eliminates the need for dedicated servers and complex installation while unlocking a scalable deployment from small store formats to large hypermarkets. The hardware-agnostic technology works with existing camera infrastructure – whether built-in, off-the-shelf or standard security cameras above the self-checkout station – eliminating the need for purpose-built AI hardware to maintain a low total cost of ownership.

With privacy at its core, it processes video on the station and is designed to support retailers in meeting their GDPR obligations. Analysing items and transaction activity only means no facial or biometric identification is required, avoiding by design the practices the EU AI Act prohibits, and is all backed by Scandit’s ISO 27001 certification.

With this announcement and Scandit’s launch of age verified self-checkout in July (iotm2mcouncil.org/iot-library/news/smart-retail-news/scandit-speeds-age-verification-at-self-checkout/), the two biggest friction points that require staff involvement are reduced. Age-verified self-checkout enables retailers to automate and verify self-checkout purchases of age-restricted items to reduce queues, remove avoidable staff intervention and improve compliance.

With both combined, Scandit (www.scandit.com) provides a more autonomous self-checkout model that supports age-verification compliance and protects against loss, freeing up staff for higher value work instead of routine checks.

Scandit’s store intelligence (www.scandit.com/solutions/store-intelligence-platform/) portfolio also encompasses on-shelf availability via ShelfView, price and planogram compliance, expiry date management and markdown, and intelligent in-store picking. Retailers can turn store data into real-time inventory visibility and prioritise actions to protect profit and reduce operational risk.