Like many restaurant owners, Andrew Valkanoff hands out bonuses to employees who’ve done a good job. But at five of his Dairy Queen franchises across North Carolina, those bonuses are determined by AI.
The AI system, called Riley, collects streams of video and audio data to assess workers’ performance, and then assigns bonuses to those who can sell more. Valkanoff installed the system, which is developed by Rochester-based surveillance company Hoptix, less than a year ago with the hopes that it would help increase sales at a time when margins were shrinking and food and labor costs were skyrocketing.
Thanks to Riley, he was able to get a 3% uptick in sales, Valkanoff said. “Every nickel right now in my business matters,” he said.
Riley, installed at about 100 stores across the country including KFC and Taco Bell franchisees, ingests data including workers’ conversations with customers and uses AI to detect whether and how often employees have tried to “upsell” (offered add-ons or extra toppings), “upsize” (offered a larger size) or tried to sign a customer up for a loyalty program. Employees who make the most suggestions and whose suggestions get converted into sales receive cash bonuses based on a scorecard generated by Hoptix's AI system. The software also tracks how quickly meals are prepared (what are the busiest times at a store and what causes long lines at drive-through) as well as how much food is wasted.
The tool is intended to be used as a coaching platform, helping managers identify and focus on individual employees whose performance might not be up to the mark and might require further training, Hoptix founder and CEO Ken Bianchi told Forbes.
“When you break this down to individual employees, you can actually start to see who's converting the most, who's not converting the most, why they are converting the most,” Bianchi said. “And now you've created a training platform where you can look at the best of the best and what they're doing. The gamification that's come from this has blown me away.”
Some experts are concerned that these kinds of AI tools can be used as an excuse to hold workers to unfair productivity standards. “A lot of times what gets called productivity issues, especially in fast-food contexts, are often a result of severe understaffing,” Alexandra Mateescu, a researcher at nonprofit research group Data and Society told Forbes. “There's a concern that such surveillance is going to further squeeze workers in a context where there's hardly anything left to squeeze.”
That hasn’t stopped the fast-food industry from wholeheartedly embracing AI-based surveillance tools that record and judge workers’ efficiency. In 2019, Domino’s rolled out the DOM Pizza Checker across all its stores in Australia and New Zealand. The checker was an overhead camera programmed with AI and machine learning that scanned millions of pizzas to ensure workers had added the correct toppings and distributed them evenly. A Portland area franchise of Outback Steakhouse, a chain of Australian-themed casual dining restaurants, also used artificial intelligence to monitor how fast its food is served and how frequently a server tends to tables. Fast food chains have adopted other less technical ways of gauging employees’ performance too, like ratings from customers that then dictate which and how many shifts workers are assigned and can even get them fired.
Bianchi, 48, was introduced to the surveillance industry 20 years ago when he founded video surveillance provider Omni Security. In 2018, Hoptix spun out of Omni Security after the entrepreneur realized that while business owners were able to record employees’ actions, they didn’t have the time to review the footage and gain any meaningful insights from it. A hobbyist baseball player himself, Bianchi was inspired by the sport’s emphasis on data, where teams have benefited from detailed metrics of each player’s performance.
“What we measure succeeds,” he said. “We're R&D-ing. And we call that ripping off and duplicating what has worked for decades in other industries, we're just doing the same thing with other physical performance-based industries, starting with restaurants.”
Riley has some checks in place to make sure that the audio transcribed by the AI is accurate and that facial recognition is identifying the correct employee in an interaction (the company says its accuracy on these points is between 96% and 99%). For the times when the AI goes wrong, the employees’ scorecard also links back to the video camera footage for a store manager to manually go through.
But some employees need convincing, Dairy Queen store operator Valkanoff said. “We’ve never run our restaurants from an eye in the sky,” he said, so he made sure to answer employees’ questions about “whether we’re watching everybody’s moves or not,” he said. “But as they start to see the impacts of this on their checks and their names going up on the board as a weekly winner, all of that starts to go away.”