AI for Theft Prevention and Process Adherence – with Alan O’Herlihy from Everseen

Daniel Faggella

Daniel Faggella is Head of Research at Emerj. Called upon by the United Nations, World Bank, INTERPOL, and leading enterprises, Daniel is a globally sought-after expert on the competitive strategy implications of AI for business and government leaders.

AI for Theft Prevention and Process Adherence - with Alan from Everseen

Episode summary: In this episode, we speak with Alan O’Herlihy, Founder and CEO of Ireland-based Everseen. Alan speaks to us about how machine vision systems can be used to detect theft or mistakes at a checkout counter (including forgetting to scan items, customers intentionally hiding items, and more). Alan not only explains where these technologies are in use today, but he also breaks down some of his own predictions about what these computer vision systems might make possible in the workplace of tomorrow.

Guest: Alan O’Herlihy, Founder and CEO of Everseen

Expertise: Machine vision for theft prevention and process adherence, retail

Brief recognition: Alan previously studied business and software engineering at University College Cork in Ireland. He has run Everseen full time since it’s founding in 2008.

(For readers with a broad interest in security applications for AI and machine vision may be interested in our full article on the security applications of artificial intelligence.)

Big Idea:

Solving unique machine vision problems (such as detecting certain kinds of theft), isn’t about the best algorithms – it’s about collecting huge amounts of visual data.

Alan frames his vision of the future of machine vision succinctly in our interview:

“AI is going to be a commodity, the land grab is access to the data. If you look at LinkedIn… someone could build LinkedIn in about 6 months. The value isn’t the software, it’s the data inside it.”

If retailers of the future want to detect theft with cameras, they’ll need more than the best cameras and the best machine vision algorithms. They’ll need tens of thousands of labeled examples of that theft behavior, captured in many, many video files. Only this kind of information (videos with many permutations of this theft behavior at the checkout counter) – says Alan – will

Companies who want to be on the cutting edge of machine vision should consider firm how they’ll stay ahead in collecting visual data at scale. Just like Google has an advantage in improving it’s search algorithm because of how many searches go through its system, leading machine vision companies in the future will be the companies with the most visual data to access and make sense of.

Interview Highlights with Everseen’s Alan O’Herlihy

Listed below are some of the main questions that were posed to Alan throughout the interview. Listeners can use the embedded podcast player (at the top of this post) to jump ahead to sections that they might be interested in:

  • (3:15) How does a machine vision system “learn” how to detect theft in a retail environment? What is involved in training such a system?
  • (8:15) How do you label the video data to train machine vision systems to detect specific events?
  • (13:15) Where does this kind of machine vision technology apply itself to the future of business? How does it extend beyond just theft prevention?
  • (16:50) How can machine vision be applied to help improve existing business processes (in warehouses, retail environments, etc) by observing existing human work processes?


Header image credit:/ck/ – Food & Cooking

Stay Ahead of the AI Curve

Discover the critical AI trends and applications that separate winners from losers in the future of business.

Sign up for the 'AI Advantage' newsletter: