Vision AI Is Taking Hold in Retail for Store Security and Analytics

Smart systems can enhance security, prevent shoplifting by repeat offenders, and gather invaluable anonymized data to inform marketing and store planning decisions.
Published: April 23, 2023

Ed. Note: This article was previously posted at Retail TouchPoints.

It’s hard to believe that in the same year cutting edge technologies like augmented reality and autonomous shopping have exploded in retail, shoplifting is still aching the minds of executives. Big box leaders like Target and Walmart have recently revealed the sizeable impact retail theft has had on 2022 revenues, with Walmart’s CEO even threatening to close stores or be forced to raise prices on consumers if things don’t change soon.

By the numbers, retailers experienced a startling 26.5% increase in organized retail crime incidents in 2021, and soon to come 2022 figures are expected to continue the trend. While there is no perfect technology out there today to end theft, retailers are meaningfully enhancing existing store systems with AI to better interpret what they’re seeing. These smart systems can help enhance security and prevent shoplifting occurrences by repeat offenders, as well as gather invaluable anonymized data to inform marketing and store planning decisions.

Preventing Repeat Shoplifting

Terry Schulenburg

Terry Schulenburg

Camera systems are the base of any retailer’s in-store security and their presence alone deters a portion of would-be shoplifters. However, for the rest, cameras alone do little to end a shoplifting incident. Unless security personnel are actively monitoring the individual out of all of the store cameras as the theft is happening, shoplifters are able to exit easily and often return to the same store to steal again. Cameras equipped with AI facial recognition can flag a known shoplifter anytime one enters the store. Once flagged, security personnel can keep a closer eye on the individual or escort them out if they’re banned from the premises.

It’s important to note that the technology cannot flag the act of shoplifting itself, or an individual shoplifter who is not known or has not been apprehended in the past. It simply acts as a heads-up to store security when a known shoplifter does enter, so they can pay closer attention. With repeat offenders accounting for 40.1% of all incidents where a suspect was identified, AI-powered security acts as an effective first step toward reducing the impact of repeat shoplifting.

Informing an Enhanced Store Experience

KPMG recently reported that retail leaders are predicting AI will have its biggest impact on the industry in the form of customer intelligence over the next two years. While reducing the impact of shoplifting on revenues is a useful first step for the technology, many agree its true potential is in better informing retailers of the habits of their shoppers to create a personalized store experience.

While AI-enhanced theft prevention uses facial recognition to match known shoplifters, vision AI for analytics dispenses with the need for any kind of matching database or recognition abilities. By integrating a vision AI system within a store’s existing camera network, retailers can gain insights into what kinds of shoppers are visiting the store, at what times, how long they’re staying, what sections they’re shopping in, whether or not they’re making purchases and more. This data can inform decisions to adjust store setup to streamline the movement of customers to priority areas, reorganize product placements, influence in-store marketing or even modify the type of music or lighting.

Industry and Consumer Readiness to Adopt

Some retailers today are using Bluetooth beacons to pick up in-store traffic data, however it should be noted that this method is far less private, connecting an individual in-store to their account or phone data. Meanwhile, vision AI systems for analytics have no idea who an individual shopper is, only assessing their general age range, gender and sometimes facial emotion.

The same KPMG study found that 90% of retail business leaders believe their employees are prepared and have the skills for AI adoption, and 53% said COVID-19 increased their company’s pace of adoption. In the case of store camera systems, there is already a framework in place to improve security and enhance customer experience based on who is coming into the store. AI is simply empowering these systems to use readily available information more intelligently, to provide actionable insights for executives.

Terry Schulenburg is a VP at CyberLink with 35-plus years of experience in the technology space, including roles at Blackboard, Genetec, Apple and more.