British frozen-food retailer Iceland says it has cut losses from shoplifting and inventory shrinkage by 80 percent after rolling out SAI's AI-based loss-prevention platform across its stores, according to Grocery Gazette.
How the system works and its effects
The solution, called SAI One, uses machine vision and generative artificial intelligence to analyse live video streams from existing in‑store camera systems. The platform monitors checkouts, aisles and entrances, and is designed to detect multiple types of loss, including items removed without being scanned, partial payments, leaving without paying, concealing products, on-premises consumption and suspicious shelf clearances.
When a suspected incident is detected, the system sends an alert to staff handheld devices within eight seconds, attaching photos and video clips. Iceland reports that false alarms have fallen to roughly half their previous level, which has increased staff confidence in the system: employees now respond to 90 percent of alerts.
Staff time, operations and financial impact
Iceland operates more than 1,000 stores in the UK, including about 700 city‑centre outlets. The retailer previously estimated annual losses from shoplifting at around £20 million. Richard Walker, Iceland's chief executive, warned that store theft undermines the company's profit margins and limits its ability to keep prices low and raise employee pay.
Andy Edwards, Iceland's head of loss prevention and corporate audit, said the technology produced meaningful improvements particularly in the 100 worst-performing stores; those locations account for roughly one third of total shrinkage. Edwards said the company aimed to make Iceland a less attractive target for thieves, which required a more proactive solution.
The company also said the system freed up staff time, allowing colleagues to spend more time on store operations and customer service.
Development plans and consideration of facial recognition
Iceland is working on further developments for the system: planned features include an alert for repeat visits that flags if a previously caught individual returns to a store, and sharing notifications across nearby stores.
The retailer is also exploring the use of facial-recognition technology as part of its crime-prevention strategy. SAI stated that if facial recognition were introduced, strict safeguards would be in place: images of innocent shoppers would not be stored, customers barred from entry could view the evidence against them, and penalties would be proportionate to previous offences.
Summary
Iceland's deployment of SAI One has substantially reduced losses from shoplifting, lowered false alarms and increased staff response rates. The change is measurable against the retailer's earlier estimate of around £20 million in annual theft-related losses, and Iceland says it will continue to refine the system while assessing ethical and legal safeguards for technologies such as facial recognition.
An AI assistant contributed to preparing this article; a journalist edited and verified the final content.
