NEXT ICON: LEXIUS

Shoplifting emerged as one of the fastest-growing challenges since the COVID-19 pandemic, leading to tremendous losses for major retail chains. Jumbo, for instance, incurs annual losses of €100 million in revenue due to shoplifting. How do we ensure that the problem is contained without becoming entrenched in a Big Brother-like society? David Elskamp provides a solution to this problem with his high-performance AI software Lexius.

This story was first published in the Next Icons Magazine. You can read the entire magazine here.

You developed software for real-time shoplifting detection. How does it work?

David: “Our AI software is an add-on to a store's existing camera system and analyzes shoppers' behavior and movements to detect potential theft. We noticed that shoplifting was one of the fastest-growing challenges after COVID-19. This is causing large retail chains to incur tremendous shrinkage. Proper monitoring is important in that respect. We are therefore filling that gap, and our mission is to implement enterprise-grade AI shoplifting detection software on a global scale. We focus purely on behavior. This includes suspicious actions such as someone tucking something into a coat or bag."

Ethnic profiling is still a common problem in AI models today. So your AI model doesn't take skin color into account, then?

David: “This is where we truly stand out. Everyone is equal within our system. You have to control AI appropriately, which is why we prioritize privacy and equality. Skin color, race, gender, or a person's clothing is irrelevant to us. We only observe movement, which makes us innovative compared to other solutions. You can resolve problems with proper AI models. We also see similar startups working with older AI models. Those also trigger many false notifications, such as a person putting a phone in their pocket, or walking around a store for a long time. Behavioral recognition as we're approaching it is state-of-the-art."

Is that also the reason why, after just a year, you are already sitting around the table with several large retail chains both in the Netherlands and abroad?

David: “That’s one of the reasons, yes. I am an entrepreneur at heart, but I don't believe in entrepreneurship if you're not contributing something to the world. At 14, I started a media agency, so I already have some experience in cameras and entrepreneurship. I also wrote my thesis on recognizing different types of crimes in cities during my bachelor's degree in Computer Science. I brought that unique blend of experience into Lexius. I am not doing this alone, as we have an international team with Ivy League talent. As a result, we have access to top universities like UC Berkeley, MIT, and Stanford. This fusion of European and American talent means that we were able to construct an AI model leveraging the latest research with extensive knowledge and a large network. That is what we owe our success to up to this point."

That sounds promising. What else can the world expect from Lexius?

David: “Currently, we have a waiting list that exploded when we opened it. Without any marketing or promotion, we now have a list of retailers with more than 4,000 stores under management that are interested in buying our product. This is quite extraordinary, but illustrates that the problem is so widespread that store owners are desperate for a solution. If we optimize our behavioral recognition, our market will be much larger than that of shoplifting alone. Behavior recognition can then also be applied in sectors such as counter-terrorism, crowd control, port monitoring, crime detection, and many more. I would be thrilled if we can use our software to make the world a safer place, without Big Brother monitoring.”