Facial recognition technology isn’t emerging; it’s here. Every day, millions of people use their faces to unlock their phones. They use facial recognition to gain access to airports and entertainment venues, secure financial transactions, and enhance security operations.
Many retailers use face matching for life safety, investigative efficiency, and loss prevention. The tool provides unprecedented visibility to the impact of repeat offenders and analyzes their patterns. These data-driven metrics allow retailers to focus their efforts and build more cases faster with higher case values.
Recently, 300 US and UK senior retail leaders told researchers the current theft crisis is their number one business concern. They cited facial recognition as their top technology solution. Today, three top ten US retailers, more than fifty grocery banners, as well as home improvement, luxury apparel, and big box chain stores rely on FaceFirst to help prevent violence and reduce loss.
We invited two retail industry leaders to share their experiences with face-matching software. Here are their thoughtful, insightful responses about their considerations and reasons for deploying the technology.
LPM: Scott, what key factors does The Home Depot consider when developing and implementing security solutions, including computer vision technology? And how do you balance building prosecutable cases vs. immediate deterrence and stopping loss?
Scott Glenn: As a retailer, stopping the loss—and making bad actors think twice about coming into our environment—are paramount. More importantly, helping our associates and customers feel safer and actually be safer is the third leg of that stool. Safety has increasingly become our responsibility. Technology solutions will be at the forefront of that activity.
You have to be focused on the details. If we have someone who is hurting us financially and is part of a true ORC group, we are more likely to allow them a little more rope in order to prosecute the larger entity. The minute they become violent or put our associates in harm’s way, we will shut them down.
LPM: Investigators use tools such as facial recognition and license plate recognition (LPR) that provide two related layers of information to build cases. How are you approaching layered solutions for life safety and loss prevention?
Glenn: Exactly the way you suggest—as complementary tools to drive efficiency and productivity. Humans cannot realistically process all of the signals that are available to us. We need these tools with machine learning and AI processing to draw out patterns, link events, and help direct our teams to the most important workflows. Often these tools are blended with other processing tools to extract the greatest value for our teams. We expect these AI solutions to be force multipliers in the long run.
LPM: Mike, what’s driving the rapid face-matching adoption in the past few years?
Mike Lamb: Businesses can get over a bad sales quarter. But the loss of associates, the loss of customers, the loss of family members through acts of in-store violence—you don’t get over those. Before FaceFirst, we had no way to take real-time action during life safety events. Now it’s the cornerstone for store safety and asset protection programs.
LPM: You consult now for various security solution providers, including FaceFirst, and you’ve always been a leader when it comes to testing and deploying new technology. During your career, you deployed a number of computer vision solutions. What are the biggest AP concerns now?
Lamb: The world’s changing. Face matching technology helps us set up a line of defense against the senseless violence we’re seeing in the retail sector. I have never seen a better tool to help mitigate life safety risks than facial recognition.
To me, the liability of doing nothing to address active threats is far greater than any potential risk from using this technology. One risk in not using it is you have individuals who are known threats, but you have no way of knowing when they come back or to provide a rapid warning to associates. You can’t possibly expect the store personnel to recognize every bad actor, witness every theft, stop every loss. This tool lets you know when they are in your stores and pinpoint criminal activity with accuracy beyond human capabilities. In the end, it is humans who are receiving these alerts, reviewing them, and responding appropriately.
LPM: Scott, life safety and individual privacy are both important. How do you balance relative risk vs. reward when implementing AI-assisted computer vision technology?
Glenn: First, we meet with our privacy and brand teams to ensure that the problem we are trying to solve is not outweighed by customer and associate perceptions. While that process can be emotional and qualitative, it still matters. I do think the macro-environment has helped make these tools seem less scary and, importantly, they have become more accepted.
LPM: Mike, what common responses do you hear from retailers that have just started using FaceFirst?
Lamb: When retailers roll out FaceFirst, they’re always surprised by how much loss they have missed using the legacy methods. I’ve been in the industry for decades, and I was shocked by what we found.
LPM: Do you think consumers’ use of facial recognition technology for their own convenience, banking, and entertainment has helped speed adoption in retail?
Lamb: I do. I just went through the Tampa airport facial recognition; lower your head, step back, boom, you’re good to go. The phone I sign on to every day uses facial recognition. I am a staunch supporter of the technology in day-to-day life. I also think we’ll see face matching used more in schools, movie theaters, concerts, festivals, religious gatherings, and other public spaces that have been targeted for violence recently.
LPM: Scott, if an everyday honest customer expressed concern about balancing individual privacy with the use of facial recognition technology, how would you respond?
Glenn: We have had these inquiries. Our response is: We go above and beyond the legal requirements for notification, and we use the tools only to protect our associates, customers, and the shopping experience. Over the years, most customers have adapted to the thought of being observed. The ubiquitous nature of CCTV has become normal. I think most of our customers understand what we are doing and trust us to protect them and their privacy.
LPM: Mike, all technology comes with a layer of risk. With face matching, the operational risk of falsely accusing the wrong person seems to be the most common concern. How did you address that concern with your privacy teams?
Lamb: The face matching tool helps employees make informed decisions about whether to observe, provide customer service, or call for law enforcement assistance. Face matching is also an investigative superpower. It’s CCTV on steroids. We saw reports that a few early adopters misused the tool. Employees with no training used bad images, then approached innocent shoppers to accuse or ban them. So, our approach was simple. We made sure the product had stringent data governance features with limited user access, clear rules, and role-based permissions. We had a robust training program, so each user was certified. We demonstrated how we were digitally enhancing our existing BOLO list. Again, the technology is not making any decision—a person is making all decisions, just in a more real-time and better-informed manner.
LPM: Scott, many retailers already use computer vision tools such as facial recognition and LPR. Do you anticipate that these tools and LPR will become the norm within five years, just as CCTV is now?
Glenn: It certainly seems like the broader environment is softening in their approach to these tools. I’d like to think retailers will continue to deploy these tools, use them responsibly, and show how valuable they can be, and that there are no inherent civil liberty dangers.
LPM: What else is important for your peers to know about developing and implementing computer vision technology successfully?
Glenn: Do the work upfront. It’s important to get this right before you deploy because you only get one chance to make that first impression. Gather all of the appropriate stakeholders early and get them on the record. Have a robust test-and-learn measurement plan. Decide what success looks like for the total corporation, not just the AP/Safety part of the organization.
LPM: Mike, your turn: What else is important for your peers to know about developing and implementing computer vision technology successfully?
Lamb: It’s important to educate the senior leadership. This is a transformational technology that helps retailers create a best-in-class threat management program and get a better grip on mitigating organized theft and the violence associated with it. Access to the system is restricted and operationalized with multiple levels of authorization and oversight to prevent misuse.
FaceFirst can help you protect people, privacy, and profits. Get in touch at facefirst.com.