Shrinkage has long been a major problem for retailers. In 2022, it cost US retailers over $112 billion. Fraudulent returns alone accounted for $103 billion in losses in 2024—underscoring the sizable and increasingly unavoidable challenges that modern retailers face in an already uncertain economy.
Traditional deterrence methods have produced mixed results, with existing practices and technologies often overwhelmed by the growing volume and complexity of offenses. In response, many retailers are turning to new solutions that promise faster and more effective outcomes.
This shift helps explain why retailers are so invested in the rise of artificial intelligence (AI) and machine learning in loss prevention.
The Challenges of Traditional Loss Prevention
Over the years, retailers have built and refined robust systems to mitigate losses. Traditional methods—such as closed-circuit television (CCTV) surveillance, electronic article surveillance (EAS) systems, and the presence of security guards—have proven effective to a degree. For example:
-
Security guards can reduce internal theft by as much as 20 percent
-
Security cameras can reduce shoplifting by up to 50 percent
-
EAS systems can reduce total inventory shrinkage by as much as 75 percent
If these numbers are accurate, why do shrink rates continue to rise?
Retail Crime on the Rise
Inventory shrinkage is increasing because retail crimes are becoming more frequent and more severe. While the basic framework of traditional prevention still holds value, the accelerating pace and sophistication of these crimes are rendering older tools less effective.
-
Shoplifting offenses rose by 93 percent between 2019 and 2023
-
Shoplifting cost retailers 90 percent more in 2023 than in 2019
-
Seventy-three percent of retailers report increased violence among shoplifters
-
Employee dishonesty rose by 18 percent between 2021 and 2022
Retail crimes are evolving, and the bluntness of conventional tools often makes them inadequate. The issue is less about the presence of security measures and more about their precision, responsiveness, and scalability.
The solution? Enhancing these measures with tools that can process large volumes of real-time data—chief among them, AI and machine learning.
The Role of AI and Machine Learning in Loss Prevention
AI and machine learning are valuable because they can analyze vast amounts of data quickly and accurately—and, importantly, learn from past incidents.
When integrated with traditional security tools, AI allows retailers greater visibility into operations and more control over how they respond to loss events. Below are several ways AI is transforming loss prevention and prompting nearly 90 percent of retailers to explore its potential.
AI-Powered Video Surveillance
Retailers are upgrading existing security cameras with AI-powered video analytics that can detect suspicious behaviors—such as loitering, repetitive motion, or aggression—in real time. These systems can instantly alert staff via mobile notifications, improving response times.
As these systems collect more data, machine learning algorithms help distinguish between normal and abnormal behavior, reducing false positives and enhancing accuracy. With AI handling surveillance alerts, security personnel can spend more time actively protecting inventory on the floor.
Behavioral Analysis
AI also supports long-term behavioral analysis, especially helpful in detecting internal theft. By identifying patterns—such as repeated access to sensitive areas at unusual times—AI can flag potential misconduct for further investigation.
Combined with inventory records, these insights can uncover long-term loss trends. Reports generated automatically can assist security and management teams in responding more quickly and accurately than traditional manual review methods would allow.
Point-of-Sale (POS) Monitoring
AI can also be integrated with POS systems to monitor transactions in real time. This allows retailers to identify unauthorized price changes, improper discount use, or transaction anomalies.
These data points can be cross-referenced with video footage to verify suspicions. Whether loss events are accidental or intentional, AI helps surface incidents for appropriate follow-up—often before major losses occur.
The Future of AI and Machine Learning in Loss Prevention
From sales data analysis to real-time inventory tracking, the applications of AI in retail loss prevention are expanding. In 2024, nearly 90 percent of retailers reported using or evaluating AI solutions—up from 82 percent in 2023—and 35 percent planned to significantly increase their investment in the near future.
As these technologies become more advanced and accessible, wider adoption is likely. The potential of AI to measurably reduce shrinkage is increasingly clear—and the outcomes of this shift will be closely watched.
