Scandit Launches Vision AI to Curb Self-Checkout Theft

The solution employs a soft on-screen nudge to alert shoppers of errors, allowing them to self-correct before completing their purchase. This approach minimizes the need for store associate intervention, which often creates bottlenecks in high-volume stores. According to data from ECR Retail Loss, stores utilizing self-checkout systems experience loss rates 22% higher in the year following installation, often forcing staff to conduct random checks that can extend transaction times by two minutes or more.

Christian Floerkemeier, CTO and co-founder of Scandit, stated that the platform is designed to eliminate the trade-off between checkout speed and security. By utilizing existing camera infrastructure, the system avoids the heavy capital expenditure associated with legacy AI hardware. The software processes video locally at the station, a design choice intended to ensure compliance with GDPR and the EU AI Act by focusing exclusively on transaction activity rather than biometric or facial identification.

Industry analysts note that reducing infrastructure latency is critical to successful implementation. Greg Buzek, President of IHL Group, highlighted that reusing existing hardware lowers entry barriers for retailers seeking to balance loss prevention with operational efficiency. This launch follows Scandit’s July 2026 rollout of Age-Verified Self-Checkout, further automating the checkout lane by reducing the reliance on human oversight for restricted items.