In 2007, Alan O’Herlihy, who previously worked with large SAP installations and in retail, set out to help retailers minimize “shrink,” or when a store has fewer items in stock than its registered inventory. He chose computer vision as a solution to the problem and founded a company, Ever seento commercialize the technology.
Everseen, which uses computer vision to prevent theft at self-checkout counters, announced today that it has raised €65 million (~$71.32 million) in a Series A round led by Crosspoint Capital Partners, a previous investor in the Startup. The new funds bring Ireland-based Everseen’s total proceeds to nearly $90 million, which O’Herlihy says will be used to scale the startup’s venture with a “focused” roadmap.
“We are experiencing significant demand for our technology from retailers struggling with the dual impact of declining customer spending and increasing operational losses, including shrink,” said O’Herlihy. “Retail is also facing challenges such as labor shortages and labor cost inflation, making our technology even more valuable in addressing these issues.”
Contraction in particular could be a serious blow to retailers’ profits, according to O’Herlihy. In 2017, stores lost an estimated 1.33% of sales to contraction, totaling an estimated $47 billion, according to at the National Retail Federation.
Everseen uses a combination of ceiling-mounted cameras and computer vision software to – in theory – reduce point-of-sale theft in physical stores. According to O’Herlihy, Everseen’s algorithms can detect and track objects of interest (e.g. SKUs), analyze how they interact and recognize “actions of interest” performed by customers and sales associates.
In addition to theft, Everseen claims to be able to “know” when items on a shelf are running low and “locate processes that need immediate attention to help staff resolve issues, improve trends, and reduce variances.” The platform, which processes videos of hundreds of millions of products and tens of millions of customer interactions every day, can connect to a retailer’s existing tools, such as an order management system, to provide insights and near-real-time analytics.
“All of these elements serve as inputs, allowing our solution to ‘prompt’ a customer to self-correct or instruct a store associate to engage and help the customer in question,” explains O’Herlihy. “Our goal is to stop and recover loss, empower the retailer to act, foster great customer interactions and create fluid processes, all while improving the overall customer experience and positively impacting bottom line.”
Everseen has not always succeeded in this mission. Employees at Walmart, once a major customer of Everseen, told Determined in a 2020 that the system often misidentified unpleasant behavior as theft and did not stop actual instances of theft.
In response to the allegations, Walmart said it had made “significant improvements” to its Everseen system, resulting in fewer alerts overall. But the relationship between the two companies deteriorated soon after. Everseen sued Walmart, alleging that the retailer appropriated the Irish company’s technology and then built its own product similar to Everseen’s. (Everseen and Walmart settled in December 2021.)
It’s hard to measure the accuracy of a system without access to the backend. But history has taught us that computer vision technology – especially technology designed to deter shoplifting – is prone to bias and other shortcomings.
Consider an algorithm trained to spot “suspicious” customer activity. If the dataset used to train it was unbalanced, for example an overwhelming amount of footage of black customers stealing, it would likely highlight the overrepresented customers more than others.
In addition, some AI-powered anti-theft solutions are explicitly designed to detect, among other things, physical characteristics of on-track shoplifting – patterns of limb movements. It’s a potentially problematic approach, given that disabled shoppers, among others, may have hallways that seem suspicious to an algorithm trained on video from able-bodied shoppers.
But assuming for a moment that Everseen is largely free of bias, there’s still the elephant in the room with any camera-based tracking system: privacy. In an email exchange, Greg Clark of Crosspoint called using Everseen’s technology to potentially capture buying intent and behavior to “sell to specific demographics” a sensitive prospect, to be sure.
I asked O’Herlihy how it handles customer data, including all the images it captures of customers and store associates. He said Everseen is deferring data retention policies to customers and – for what it’s worth – is “fully compliant” with GDPR.
Whether shoppers – or employees for that matter – trust Everseen implicitly is another question. But the potentially thorny ethical issues don’t seem to deter clients from signing up for the startup’s services.
O’Herlihy claims that Everseen counts more than half of the world’s top 15 retailers among its customers, with deployments in more than 6,000 stores and more than 80,000 checkouts.
“The adoption rate of this transformational technology increased during the pandemic as retailers looked for different ways to sell and shoppers looked for different ways to buy,” said O’Herlihy. “In terms of technology spending, we’ve seen a reallocation of budgets as challenges for retailers evolve and addressing shrink is seen as a top industry priority…Everseen aligns seamlessly with current trends.”
In a general sense, it is true that retailers are embracing or at least showing an interest in AI. a recent KPMG research found that 90% of retail leaders believe their employees are prepared and have the skills for AI adoption, while 53% agree that the pandemic has increased their company’s adoption rate.
In the future, Everseen will – no doubt under pressure from rivals such as AI watchman and VeelEye – plans to expand its technology into sectors beyond retail, such as supply chain and manufacturing. The startup currently employs about 1,000 people across its headquarters in Cork and hubs in the US, Barcelona, India and Australia and elsewhere.
“Beginning with retail allowed Everseen to develop both a foundation and library of computer vision AI use cases relevant to other adjacent industries,” said O’Herlihy. “Computer vision solutions are currently very isolated and focused on solving specific problems. We see an increasing demand for our platform as customers try to solve other problems in the shopping area.”