CHICAGO -- Step inside a Loop Neighborhood convenience store in San Francisco and you might meet an unusual employee. She won’t say hi when you enter the shop, ask you about your day or remember your name. But she can make you 50 entree-sized salads.
Sally the Salad Robot offers 12- to 14-ounce salads, either premade or customized with a choice of up to 22 ingredients. Operators can pick from about 115 ingredients in programming her selection. Sally is not limited to salads, either. She can also make quinoa grain bowls, breakfast cereal, yogurt bowls and snacks such as hummus and pita chips.
At another Loop store, a cone-shaped white robot—a cross between R2-D2 and a NASA space shuttle—patrols the forecourt, scanning license plates and recording activity around the parking lot. And at Dodger Stadium in Los Angeles, a foodservice robot called Flippy flips hamburgers on a grill, the kind of tedious task that can try employees’ patience.
Some may see these robots as taking a job from a human. But Alison Acerra, vice president of nutrition for Redwood City, Calif.-based Chowbotics, which manufactures Sally, insists that is not the case
“We see Sally as a job creator. Similar to servicing a salad bar, operators will need employees to prep ingredients, fill machines and clean,” she says. Essentially, human employees will be reassigned to jobs for which they are best suited. And today, the danger is not from robots taking jobs—it is from the lack of humans to fill these jobs.
National unemployment sat at 4% as of June 2018, according to the Bureau of Labor Statistics (BLS). And that 4% is based on a smaller labor pool: The labor force participation rate has been around 63% since 2013, a level not seen since the late 1970s. It now takes on average 31 days for an employer to fill a new job, up from 23 days in 2006, according to Deutsche Bank.
“We’re entering the tightest labor market that any of us has ever seen,” economist Bill Conerly said at the NACS 2018 State of the Industry Summit. The working-age population will be lower from 2020 to 2030 than it was during the Civil War, when about 2% of the population died on the battlefield, according to the American Battlefield Trust.
Changes in the country’s labor picture are also affecting industries supporting convenience retail, namely trucking. Since 2015, trucking hires have averaged just 1% annual growth, according to the BLS.
“You have a lot of baby boomers retiring [from trucking], and you have a massive gap in the number of people retiring and then finding new people to fill those positions,” says Jack Karsten, a research analyst with the Brookings Institute, a public policy organization based in Washington, D.C. “This is going to be a huge problem because almost everything that arrives in a store got there by truck, so if we have massive labor shortages in trucking, that could raise the price of pretty much any retail good that people buy.”
Scott Oakes has seen retailers’ desperation firsthand. At Gilbarco Veeder-Root’s recent Retail & Fuel Technology Conference, Oakes, manager of North American point of sale and foodservice solutions, gave a presentation on a self-checkout system undergoing testing with c-store retailer R.L. Vallee, Saint Albans, Vt. The self-checkout platform will not be available commercially until early 2019, but many attending retailers wanted it now.
“Some states mentioned to me have virtually 0% unemployment,” Oakes says. In March 2018, eight states reached their lowest unemployment rate since the state unemployment data series began in January 1976, including Hawaii at 2.1% and New Hampshire and North Dakota, each at 2.6%, according to the BLS.
“[Retailers] said they couldn’t find anybody to hire to work in their store,” Oakes says. “So instead of [self-checkout] just being a cost savings, it’s about viability of your business, because they can’t find people to work in the store and automation is the only answer.”
While automation, artificial intelligence (AI) and robotics seem futuristic, retail and restaurants are already brimming with it. For instance, the Coca-Cola Freestyle machine, a smart fountain dispenser with more than 100 flavors, monitors the flavor combinations customers choose and uses that data to suggest new Coke products. And many types of equipment, from coffee makers to gas pumps, are being integrated with internet of things technology, allowing them to communicate and respond remotely.
As robotics and AI become more sophisticated, they will play a growing role in delivery, through drones. But these will likely not be the flying variety. Gray Taylor, executive director of tech standards group Conexxus, Alexandria, Va., says the real future is in terrestrial drones, which come in a variety of shapes and sizes, from sidewalk-bound rolling coolers from London-based Starship Technologies to the almost sedan-size delivery vehicles from Nuro, Mountain View, Calif., which Cincinnati-based Kroger has tapped for its same-day delivery test.
During the test, customers will be able to place same-day delivery orders through Kroger’s online ClickList ordering system and Nuro’s app. Nuro’s fleet of self-driving delivery pods will then delivery the groceries. Kroger said it will announce the pilot market for the test soon, and it expects to begin the test this fall. If the test is successful and Nuro vehicles begin delivering from more of Kroger’s 2,800 stores spread across 35 states, it could be a substantial step forward for autonomous delivery.
“That one’s got legs,” says Taylor of Nuro’s concept. “It’s going to be easy out of the gate, cheap to operate, and it’s going to require a really rudimentary level of intelligence for it to function. Because of its mass, it’s not going to be seen as one of those things that can cause a huge accident if it malfunctions. It won’t kill you if it hits you.”
Robots also have tremendous application in automating labor-intensive foodservice tasks such as flipping burgers, making pizza or brewing cappuccinos. Flippy the burger-flipping robot from Pasadena, Calif.-based Miso Robotics has been cooking fried chicken tenders and tater tots at Dodger Stadium. The robot is essentially a robotic arm with cameras and sensors that allow it to work with food and human employees. The stadium is employing Flippy for time-consuming, repetitive tasks, allowing human employees more time to finish dishes and interact with customers. According to Miso Robotics, its autonomous robotic kitchen assistants start at $60,000 and include a 20% annual fee for “AI continuous learning” and maintenance.
Sally the Salad Robot and Flippy aren’t the only robots lending a hand. At one Loop Neighborhood store in San Francisco, a security robot—reportedly a K5, manufactured by Mountain View, Calif.-based Knightscope—trundles around the parking lot as a crime deterrent. It scans license plates, records activity around the store and sounds alarms to deter loitering. The machine-as-a-service operation fee is about $7 an hour. K5 has also been the subject of many selfies taken by curious customers.
Automation goes beyond buying a fancy robot. There are more economical ways to start automating store tasks, says Taylor—in particular, robotic process automation (RPA). It involves teaching computer software to complete a repeatable, data-intensive task normally done by a human. Most often, this starts in the back office.
“It’s not about automated pizza making. It’s really going to come in the supply-chain side—that’s where we’re going to see the low-hanging fruit,” he says. “Most people don’t really know how much time a manager is spending on building an order. So is there a way that a machine can go in and build that order?”
The first step in setting up RPA is for c-store company leadership to discuss how to eliminate redundancies in the labor model related to processes. “And don’t be self-selecting,” Taylor says. “Do it with an open mind, as if everything is attainable.”
RPA could save c-stores time in often overlooked areas. For example, say a few sales executives from a large retailer just returned from CSP’s Outlook Leadership conference. They made plenty of new connections and collected a book’s worth of contact information, but someone has to input that data into the customer relationship management software. The sales execs might throw this tedious data entry work at a lower-level employee, but if they have access to RPA, the software could accomplish the task far faster than any human being could. Freeing up the lower-level employee’s time may allow them to interact more with customers.
“Ultimately it is about being able to either reduce hours or reallocate people from standing behind the counter.”
Taylor says RPA used to be a manual process in which an employee had to take the time to observe another employee and investigate their productivity. Today, there are AI engines that can follow and mimic a person completing a simple manual process. “Are you ready to do this in the 20-store-chain range? Probably not yet,” Taylor says, “but there is a lot of information coming back that’s showing that this type of technology is being very commoditized and democratized.”
For an even higher investment, automated solutions based on AI or machine learning can improve processes. Austin, Texas-based Revionics, for instance, uses machine learning to analyze data on a chain’s past promotions, prices and layout. It then takes those findings and suggests strategies to optimize each. “It’s constantly looking at your history, and it’s gaming itself,” Taylor says. “It’s going through and asking what promotion works best or the least well with this or that promotion. It starts building a profile of your individual stores and demographics.”
For retailers looking to improve operations, San Francisco-based Zenput automates task management, even across chains with dozens of stores. Zenput connects every employee across multiple locations to the main office through a mobile app. Once employees complete a task assigned to them through the app, they can mark the task complete or take a picture of the completed task with their mobile device.
“Otherwise, you’re kind of like this human gardener,” Taylor says. “Then you invariably get to the last day of the week and you haven’t done everything to the extent you need to.”
Considering the squeezed labor market, automating in-store tasks can also help reduce a store’s total hours worked, says Skip Potter, IT manager for R.L. Vallee. That is part of the reason the c-store retailer was interested in partnering with Gilbarco on self-checkout stations.
“Ultimately it is about being able to either reduce hours or reallocate people from standing behind the counter,” Potter says. “Instead of having two [employees] there, you can rotate one out to your coffee bar or out onto your floor. You can reallocate where you’re positioning people and, at the end of the day, ultimately reducing hours out of the budget.”
Eliminating 10 hours of work through automation is a great goal, Potter says, but even cutting four to six hours a day is a win. Taylor of Conexxus agrees. The goal is cutting “not staff, necessarily, but staff time,” he says.
Since starting his own self-checkout test, Potter of R.L. Vallee has become something of a self-checkout stalker, constantly observing the system and its effect on consumer behavior. From what he’s seen, anyone can be a self-checkout user.
“A person uses it, they buy into it, and then they’re going to be a self-checkout person,” he says. “And you see that more and more. You see that in grocery stores, you see that in convenience—even McDonald’s. I actually see lines at self-checkouts when there are attended registers.”
Some stores are more optimized for self-checkout than others. Potter and Oakes agree that careful planning and testing is needed to make sure self-checkout works in a chain before it is implemented.
“One of the key things for self-checkout is signage, being able to direct the person to where they need to be. Positioning is certainly important, but I think that signage is equally as important,” Potter says. At one store with self-checkout, R.L. Vallee hung a large green sign with a red hand pointing to the station that reads, “In a Hurry? Use Self-Checkout.” You can’t miss it.
The teams at R.L. Vallee and Gilbarco also discovered that the self-checkout process was more seamless if the stations accepted payment only through credit or debit cards or mobile payment. Cash is best left for humans to handle. The chain’s self-checkout stations can pause a transaction and continue on a manned point-of-sale in case a customer is unable to check out on their own or is purchasing an age-restricted item.
R.L. Vallee and Gilbarco are putting more self-checkout stations in more locations, albeit carefully. The chain is installing custom checkout counters in a new store with fake fronts. Once R.L. Vallee is ready to launch self-checkout at that location, it can replace the fake counter fronts with the self-checkout stations.
QuickChek, Whitehouse Station, N.J., pioneered the c-store industry’s use of self-checkout stations in one of its New Jersey stores in 2008. John Schaninger, former vice president of sales and marketing for QuickChek, and now owner of consultancy The Schaninger Group, Bucks County, Pa., was with the c-store chain during its successful rollout of the technology. He saw sites that needed four cashiers at peak times fall to two with the addition of self-checkout stations. Extra cashiers were moved to the sales floor to directly serve customers.
“Much of this technology can give a nod, sometimes a large nod, to reducing hours at the store level, but the focus must be on pleasing customers daily,” Schaninger says. “If we use technology to so satisfy customers that traffic and sales increase, this should result in additional profits covering costs.”
“[QuickChek] customers love their self-checkout because nobody ever stands in line,” Taylor says. “They have an ambassador
who’s walking through all the four self-checkout lanes and greeting the customer and so forth, so there are measures that you can take between going staffless and where we are now, which is full-service.”
Hope for the Future
The forces that will forever alter the retail landscape will also alter our society. Taylor says we are approaching a world in which AI engines predict our needs and make purchases for us. He points to Google’s Psychic Pizza project, which would, for example, automatically deliver a soda to a customer while they are having a pizza, knowing that said customer enjoys a soda with their pizza.
AI making purchasing decisions for us might sound creepy now, but Taylor predicts we will grow accustomed to the idea.
“The argument will be that it is a reflection of your free will,” he says. “You always get a Coke when you order a pizza, so the AI engine just took the liberty of passing that along. By the way, if you don’t want it, we’ll send it back. That’s when you get used to it.” Once decisions are made and are carried out by an army of robots and AI, both the industry and the nation will be forced to re-evaluate the way everyone thinks of work, employment and benefits.
As positions such as checkout become more automated and employees are forced to spend more time on the store floor with customers, c-stores will have to hire differently.
“You’re going to start hiring people who are great with people rather than people who have great cashier or lifting or quantitative skills,” Taylor says.
“[QuickChek] customers love their self-checkout because nobody ever stands in line.”
Karsten of the Brookings Institute says society will have to keep an open mind when approaching this future.
“One idea is to radically expand our definition of work,” Karsten says. “Right now, we think of it as paid labor. You do a task and you are compensated for that task. But what if we thought about it as doing volunteer work in your community or as taking care of a child or an elderly relative? We could expand our definition of work to cover more activities and distribute benefits based on those things and not just paid employment.”
Karsten believes that the transition to a new definition of work and labor will be difficult. “If you think back to the Industrial Revolution, there was a lot of social strife around the movement of people away from farms and into factories. That was a lot of social upheaval,” he says.
But Karsten does not think the mass adoption of AI into the workforce will be disastrous, either. “Back when ATMs were introduced, there was this worry that they’d put all bank tellers out of business, but if you look at employment numbers from bank tellers, it’s actually gone up since ATMs were introduced,” he says.
The number of bank tellers continued to increase because the technology allowed banks to open more locations with fewer tellers in each, while some of the tellers moved into higher-value work such as selling loans.
As workers continue to be phased out through automation and the very idea of c-stores become obsolete as technology advances, it will be up to the industry to reimagine its value and place in the future.
“Sometimes we can be so conservative that we’re the last ones to embrace an innovation because, well, that’s just not the way we do it,” says Taylor of Conexxus. “Well, that’s the way bookstores used to think. That’s the way taxicab companies used to think. That’s the way the travel agencies used to think. We need to be more healthfully paranoid about what can happen to us overnight.”
Robot History in Retail
Members help make our journalism possible. Become a CSP member today and unlock exclusive benefits, including unlimited access to all of our content. Sign up here.