Retailers Use AI and Edge Compute to Combat Inventory 'Shrinkage'

- Computer vision systems can identify specific behaviors that may indicate theft, such as customers lingering in unusual areas, concealing items, or bypassing checkout counters. These AI-powered cameras can also detect when an item is placed in a bag without being scanned at self-checkout and immediately alert staff. Some systems can even recognize and flag known shoplifters upon entry by using facial recognition technology. - Edge computing architecture allows for real-time processing of video analytics on-site, which significantly reduces the delays associated with sending data to the cloud for analysis. This enables instant alerts for security personnel to respond to potential incidents as they happen. This localized processing also enhances data security by minimizing the transfer of sensitive customer information. - In 2022, inventory shrinkage accounted for $112.1 billion in losses for the retail industry, an increase from $93.9 billion in 2021. These losses are not just from external theft but also include employee theft, supply chain fraud, and administrative errors, which together account for nearly two-thirds of all shrinkage. - Beyond video surveillance, AI is being used to analyze a wide range of data to detect and prevent fraud. Machine learning algorithms can identify subtle, fraudulent patterns in transaction data with high accuracy, and natural language processing can analyze customer communications to flag potential scams. - RFID technology offers a significant boost to inventory accuracy, with some retailers seeing an improvement from 65% to as high as 95%. This is because RFID tags don't require a direct line of sight to be read, allowing for rapid and frequent inventory counts with handheld scanners or automated read points at key locations like dock doors. - The impact of shrinkage extends beyond direct financial loss, leading to increased operational costs for security and inventory management. To compensate for these losses, retailers may raise prices, which can negatively affect customer satisfaction and loyalty. In some cases, high levels of theft have forced stores to close or reduce their operating hours. - AI algorithms can enhance demand forecasting by analyzing historical sales data, real-time demand signals, and even external factors like weather and social media trends. This leads to more optimized inventory levels, reducing the likelihood of both stockouts and overstock situations, which can tie up capital and lead to losses. - While AI offers powerful tools for loss prevention, it also presents new risks, as fraudsters can use it to carry out more sophisticated and large-scale attacks. This has led to a notable increase in fraudulent online traffic for some retailers. As a result, only a small percentage of retailers feel adequately prepared to handle the emerging threat of AI-enabled fraud.