Abrupt changes in how people shop—brought about with sudden lethality during the pandemic—is dovetailing with a march toward automation and artificial intelligence likely to infiltrate just about every aspect of doing retail food business, from the warehouse to the refrigerators in shoppers’ homes.
These changes are further quickening evolution in the industry, where efficiency matters more than ever, in part to afford the experiential, omnichannel and price-competitive offerings today’s shoppers demand. It’s also a state of change likely to stay around—a condition of the exponential nature of technologies.
“Like the simple caterpillar giving way to the complex butterfly, retail is undergoing a metamorphosis. The DNA of traditional retail is becoming digitized and mixed with accelerating and converging technologies,” Gary Hawkins, a former food retailer turned CEO of the Center for Advancing Retail and Technology (CART), writes in a newly published essay.
Where should retailers be looking for innovation? What follows are trends, in and of themselves, that are also related by common threads of artificial intelligence, machine learning and automation.
“The future of robotics in retail is very bright,” said Rian Whitton, senior analyst for robotics and automation for ABI Research, a New York strategic consulting firm, during a recent WGB webinar examining progress of store-based robotics such as automated floor cleaners and stocking robots. These units are wheeling though a small but growing number of retail outlets today but “hundreds of thousands” will come to retail sites by 2025.
Whitton noted that recent technological improvements in AI, sensors, communication and processing hardware have made mobile robotics a “readily deployable technology.” Consumers, in the meantime, have grown more accepting of machines in public spaces, which is moving robots from the factories and warehouses where they’re already widely deployed into public spaces such as retail stores and “connected smart cities.”
He predicts retailers in 2021 will test and evaluate various solutions seeking “scalability,” with heavier deployments in the years beyond.
“It’s really about generating use cases and education at this stage, so you can iron out the problems and get to profitability,” Whitton said. “The main challenge is scalability. But by 2022, we believe companies will be building up the accompanying infrastructure to enable operators scalability.”
Robots, he noted, don’t simply provide productivity in labor but when enabled with AI and machine learning, will improve over time, “even to the point where labor scarcity becomes less of an issue, and the benefits of this productivity growth will really be reached by 2025 through safety and productivity improvements.”
Already something of a showcase for tomorrow’s retail technology, two Seattle area Amazon Go stores earlier this year debuted Amazon One, a payment identification technology that allows shoppers to check out not by swiping their phone over a sensor but by waving the palm of their hand.
“The service is designed to be highly secure and uses custom-built algorithms and hardware to create a person’s unique palm signature,” said Dilip Kumar, VP of physical retail and technology for Amazon, in a blog post.
According to Amazon, it takes less than a minute for customers to sign up to use an Amazon One device. After inserting a credit card, shoppers hover their palms over the device and follow the prompts to associate that card with their unique palm signature. Shoppers have the option to enroll with just one palm or both. Once enrolled, users can hold their palm above the Amazon One device for store entry and payment.
Though seemingly less “creepy” than other forms of technology-enabled sensed identifications—such as facial recognition—biometrics may not sweep the industry right away. Some states and localities are enacting strict privacy laws. But the draw of further conveniences—and the increasing importance that interactions are as “contact-free” as possible as America fights the coronavirus pandemic—seem likely to grow in years ahead, and as Amazon noted, retail-store payment is but one potential deployment.
Josh Parks, an entrepreneur in Texas, is incorporating a similar technology as part of a developing solution to enable more efficient deliveries of groceries and essentials he’s calling Boxcart. Though still in the early stages, Parks contemplates Boxcart using biometrics—specifically, the PalmSecure palm vein authentication from manufacturer Fujitsu, which sells units as a service to enterprise organizations—as a “key” to unlocking storage lockers in dense residential buildings. These units could, in turn, provide deliverers with a more efficient way to “batch” orders.
Parks in an interview with WGB says he saw Amazon One as offering something of a novel wrinkle that incorporates the curvature of the hand in addition to palm vein patterns—which present a unique “signature” likened to a fingerprint and enabling palm biometrics like that of Fujitsu.
Anxious for solutions to one of the biggest costs and profitability impediments to omnichannel shopping, retailers have been experimenting with driverless delivery vehicles for some time, but it appears that appetite is running well ahead of local regulations.
That gap is providing an opportunity for entrepreneurs who see huge benefits to retail in semiautonomous delivery and products that are street legal right now. California-based Tortoise, for example, uses light electric vehicles to ferry abandoned ride-share scooters to popular pickup points and charging stations via remote drivers—basically, a video-game player in a Mexico call center controlling the vehicles by joystick and a screen—and is now using the very same technologies to enable an electric delivery wagon just small enough and just slow enough to legally ride on shoulders, bike lanes and sidewalks of virtually any city today.
The Tortoise vehicles can carry about 100 pounds of goods—or about eight bags of groceries—for short trips at less than 7 miles an hour, making a cost-effective solution for two-hour grocery deliveries in a tight radius of a store, Dmitry Shevelenko, a former Uber executive who co-founded the company, told WGB in an interview. Its cost is about a $1 a mile. “At a minimum, that’s 15% cheaper than your next-cheapest option,” he says, which would likely be a courier from the store or a crowdsourced driver.
Hawkins in an interview described what he calls process automation as another area of tech-enabled retailing likely to grow in coming years. This, he says, leverages AI power and software to automate processes such as marketing personalization.
“Today, a retailer chooses an audience, let’s say their ‘gold’ customers. They put together an offer pool of a few hundred items. And then, the engine figures out which offers to give to which shoppers,” he says. “If that becomes automated, what happens is the system itself surfaces what audience the retailer should be focused on, then automates the recommendations. It’s the system itself saying, ‘Here are the products, and here are the prices to promote those products to each individual customer for this campaign.’ So all the retailer has to deal with is to put a budget in and approve it or not approve it.”
The software company Birdzi, of which Hawkins is a board member, is enabling this now at companies such as Weis Markets, he says.