One of the best ways to deal with shoplifting is to make sure it doesn’t happen in the first place.
This is the goal of Paris-based AI startup Veesion, which has developed an algorithm that can recognize gestures to predict potential retail theft cases.
“I happen to have an uncle who runs and runs three supermarkets in Paris, so I know exactly what shoplifting representatives are for retailers,” co-founder Benoît Koenig told Business Insider.
Veesion said its technology has been deployed in 5,000 stores in Europe, Canada and the US. The startup recently raised a $43 million Series B funding round to promote expansion into the US.
Shoplifting vigilance has subsided somewhat over the past year as retailers and law enforcement agencies have a better grasp of the issue. According to data from AI research platform Alphasense, the term “Shrink,” a trade term for shortages, has declined significantly among major retailers’ business insider trucks.
But despite the fact that shoplifting headlines are becoming less (particularly compared to Retail’s flashy new AI capabilities), Koenig said the issue is compelling to tackle machine learning.
“It’s not appealing, but the ROI is very direct,” he said. “You arrest shoplifters, collect inventory and save money.”
One important difference between Veesion’s technology and several other visual security approaches is that they do not rely on individual tracking or physical properties that can cause bias or concerns about individual privacy.
The algorithm doesn’t care what you look like, said Benoît Koenig, co-founder of Veesion. It only cares about what you’re doing. Veesion
“The algorithm doesn’t care what people look like. It cares about how your body parts move over time,” Koenig said.
The system analyzes footage from existing security camera networks to detect people in the photograph, identify their movements, and recognize various objects such as products, carts, baskets, bags, and more.
If the movement is deemed suspicious, the video clip will be flagged and sent to the security officer, who can investigate or intervene. Security teams can update the app with additional details about whether an alert is required, whether the theft has stopped, and how much value the stolen item is.
Koenig said that over 85% of Alerts are marked as relevant to store operators using the Veesion system. He said one US client could cut losses from the health and beauty section in half in the first three months of implementation.
While many US retailers address shoplifting problems by locking items and limiting the way people shop, their approach appears to be increasingly backfiring in the form of lower sales and worsening customer experience.
“Retailers are implementing a number of security measures (that undermine the shopping experience) to protect their products from theft and keep their employees and customers safe,” the National Retail Federation said in a December report on retail crimes.
By layering the store’s existing security camera infrastructure and alerting staff to certain dangerous behavior, Veision says its technology will help create a more comfortable shopping trip.
This is a demo of alerts that store security personnel receive when Veesion’s system detects who hides an item. Veesion
Koenig said the technology also helps reduce employee theft. This estimates that industry groups estimate retailers as much as shoplifters.
Related Stories
Business Insider tells the innovative stories you want to know
Business Insider tells the innovative stories you want to know
“It has an internal deterrent effect,” Koenig said. “They know that there is AI on camera, so they’ll be careful about what they do.”
There are also retail use cases that Veesion is also exploring, including improper scans for slip-and-fall detection at self-checkout.
For now, Koenig said the technology is not effective in detecting and disrupting shoplifters. “This isn’t just a recovery of a few dollars,” he said.
