AI in Russian Stores Detects Potential Shoplifters
The System Analyzes Surveillance Footage and Flags Suspicious Behavior in Real Time
Russian company VIDAR, a developer of machine-vision systems for retail, has unveiled an AI-powered anti-theft platform that analyzes customer behavior in real time to predict and prevent shoplifting — all without using biometric data.
The neural network processes live video feeds from standard in-store cameras, meaning retailers can use the system without upgrading their existing infrastructure. By applying predictive analytics, the AI evaluates shoppers’ movements, forecasts possible thefts, and identifies irregular situations.
If a theft occurs, the system tracks the item’s movement to the checkout and cross-checks whether it was paid for. In the event of a mismatch, an instant alert is sent to security staff or directly to their mobile app.
Tracking Repeat Offenders
After an incident, the system assigns the person an anonymous ID, stored in a secure database. If the same individual appears later in any other store connected to VIDAR’s network, the system gives them a negative risk rating, prompting closer AI monitoring.
During pilot testing, the platform helped reduce theft by over 87% and improve overall operational efficiency by up to 90%.
Three large retail chains are already piloting the technology. VIDAR’s portfolio also includes modules for security and checkout monitoring, merchandising control, and visitor analytics — positioning the company at the forefront of AI-driven retail management and loss prevention in Russia.
