February 09 - 11, 2020
Miami Marriott Biscayne Bay, FL
Walmart's AI-Powered Store of the Future Is Nothing Like Amazon Go
(Image source: Walmart)
The world's biggest retailer, Walmart, is looking to change the nature of shopping as we know it. But by harnessing the power of AI at its Intelligent Retail Lab (IRL), Walmart isn't seeking to replicate the Amazon Go experience.
Recently, Amazon has been attracting a lot of publicity - and a fair amount of controversy - for its effort to revolutionize the buying process through the ultimate example of retail self-service: The Amazon Go store.
When Amazon launched its first Go store in 2018, the public lined up around the block to see the future of retail: a new experience where you could walk in, grab something off the shelf, and walk out. Sure, there were cameras on the ceiling and AI on computers tracking silently from above, but the promise was convenience through automation — maybe not The Jetsons, but a futuristic 7-Eleven for certain.
Observers wondered how others in the retail space might respond. Now, one of the online giant's principal competitors on the ground has pioneered its own form of retail revolution — but not quite in the way you'd expect.
Walmart has shared its version of the future of brick-and-mortar retail, the Intelligent Retail Lab, or "IRL" for short. Unlike Go, it doesn't feature any futuristic user experiences. There's no automated checkout or similar bells and whistles that people will post to Instagram about. Instead, IRL can track Walmart's inventory in real time with unprecedented efficiency, making sure every item on every shelf is always in stock.
Back to Basics — But With Cutting Edge Tech
For Walmart, a back to basics approach holds the key to retail's future. Rethinking the entire shopping experience, as Amazon Go has done, wasn't on the table. "It's just not a priority for us right now, as we think about it," says Mike Hanrahan, CEO of IRL (which is technically a startup within Walmart itself).
Stockout syndrome is the nightmare condition that all retailers fear — and for good reasons. Lapses in inventory result in lost sales, reduced customer satisfaction, and diminished loyalty. When customers can't get what they want at your store whenever they want it, they're also more likely to badmouth your company as being unreliable and to go elsewhere in the future.
Avoiding stockouts makes good business sense, and it helps preserve your company's reputation. To this end, the IRL store has 1,500 cameras hanging from the ceiling to ensure that when you walk up to the meat section, there's plenty in stock. "If you have really good inventory, it leads to a better managed store," says Hanrahan. And a better managed store is a more profitable one.
Using visual technology to keep an eye on what's in stock is just one avenue of best practice for inventory management. Others concern the people, processes, and technology involved in the day to day running of a store. All of these areas are being addressed at Walmart's new experimental facility.
The Intelligent Retail Lab (IRL) At Walmart
Recently, Walmart has been quietly transforming a Neighborhood Market in Levittown, New York, in line with its vision for the future of retail. With AI-enabled cameras, interactive displays, and a massive data center, Walmart's new Intelligent Retail Lab provides a unique real-world shopping environment designed to explore the possibilities which artificial intelligence can contribute to the store experience.
To create a viable testing environment, Walmart's tech incubator, Store No. 8, has placed the IRL within one of the company's busiest locations: a real store containing over 30,000 items. That said, the location is roughly half the footprint of your average suburban Walmart Supercenter.
"We've got 50,000 square feet of real retail space. The scope of what we can do operationally is so exciting," says IRL CEO Mike Hanrahan. "Technology enables us to understand so much more - in real time - about our business. When you combine all the information we're gathering in IRL with Walmart's 50-plus years of expertise in running stores, you can create really powerful experiences that improve the lives of both our customers and associates."
For the past 18 months, Hanrahan's team of 75 has been retrofitting the store with the previously mentioned 1,500 cameras (which include standard cameras and 3D depth-sensing cameras), along with various sensors on the shelves which can measure weight, count remaining inventory when an item is taken, and even use capacitive technology, much like your iPhone touchscreen works, to measure the electrical resistance of items on the shelf to see their shapes and count them up.
All of this hardware is connected by 150,000 feet of cabling - enough to scale Mt. Everest five times - and enough processing power to download three years' worth of music (27,000 hours) each second.
Educating The AI
At the Intelligent Retail Lab, a combination of cameras and real-time analytics will automatically trigger out-of-stock notifications to internal apps that alert Walmart associates when it's time to re-stock. In effect, this requires the store to automatically detect a product on the shelf, recognize specific products (including the difference between specific amounts of the same product), and compare the quantities of product on the shelf to the upcoming sales demand.
This approach to artificial intelligence incorporates several of the established best practices associated with good inventory management: conducting regular stock counts, knowing your own inventory, predicting demand, and enabling rapid responses to changing conditions. It also requires the AI itself to be powerful and smart enough to make fine distinctions between individual items and amounts of items on a shelf.
When the IRL store's hardware flipped on in October 2018, Walmart needed data with which to train its AIs — most of which operate using various types of visual recognition. This typically requires tens of thousands of examples to be fed into a system, so that the artificial intelligence can recognize the difference between an apple and a banana, for example.
So for a few months, all Walmart's technical team did was collect data. With numerous scenarios to consider, the focus was on learning from the technology, and not implementing changes to operations in haste.
Rather than a massive central system, AI training was sectioned off using a range of specialized software tools, each dedicated to a particular task. For example, there's an inventory tracker for Walmart's meat counter, which can identify every type of meat sitting on the shelf. If a customer selects an item, the AI will know exactly what is now missing and relay a text message to an associate, who can then restock it.
Little tasks like this can make an enormous difference. That tiny function of tracking meat can have a huge impact on Walmart's bottom line — and on the customer experience. For example, if you go to Walmart to get chicken breasts and discover that they're all gone, you will leave without buying them. Walmart loses a sale and you get angry for wasting your time. It's bad for everyone. Multiply that phenomenon across the 30,000 different products in the Levittown store, and it begins to scale in a meaningful way.
Having real-time inventory also allows Walmart to accurately update its website and app, which in turn helps Walmart to keep pace for quick shipping or in-store pickups with Amazon, Target, and other competitors. It all ties back to knowing what's in stock.
Walking into the Intelligent Retail Lab for the first time is both familiar and unique. There are a lot of the traditional elements you'd expect from a Walmart Neighborhood Market: associates, cash registers, and shelves stacked with thousands of products. There are also features that sit in stark contrast, such as a glass-encased data center bathed in a blue glow.
(Image source: Walmart)
This is where the computer brains behind IRL's artificial intelligence reside. The move to make it so clearly visible reflects a conscious decision on the part of Walmart's management.
Much of the controversy surrounding the tech behind the Amazon Go experience stems from its association with surveillance culture and the current atmosphere of blatant disregard of personal privacy being exhibited by so many players in the Big Tech arena. Looking to avoid these pitfalls, and in a bid to enhance the customer experience at the Intelligent Retail Lab, Walmart is making its own AI technology an active and interactive part of the in-store architecture.
"We chose right from the very start to not hide the technology," Hanrahan explains. "I think it's always good to be transparent when it comes to the world of AI."
(Image source: Walmart)
Nothing about Walmart's cameras or computers is hidden. And the store includes multiple information stations for customers to visit and understand exactly how AI makes the store tick. As customers shop, they can interact with a number of educational displays, while small educational kiosks are distributed throughout the store. A Welcome Center at the front end allows customers to dive deeper into technical specifications and common questions.
(Image source: Walmart)
Flanking the Plexiglas windows just outside the Data Center where the servers are housed are two large displays — one of which encourages participants to move around and learn how the technology reacts to body positioning. It's a fun and educational way for customers to engage with the IRL experience. Walmart even welcomes school groups and organizations like Girls Who Code to tour the store.
(Image source: Walmart)
Enhancing the Associate Experience
At the Intelligent Retail Lab, the experience for Walmart's own associates is taken into consideration, alongside that of their customers. After all, these are the personnel who must take responsibility on a daily basis for keeping those customers happy and for keeping fresh produce on the shelves.
Basic principles and best retail practices remain at the core of IRL research. Before jumping to more futuristic concepts, the IRL team is starting with real, practical solutions like the meat inventory example, as well as others like making sure shopping carts are available and registers are open.
With the technology available, the IRL research team will use real-time information to explore efficiencies that allow associates to know more precisely when to restock products, so items are available on shelves when they're needed. Because of the levels of automation that the IRL's hardware and artificial intelligence make possible, associates won't have to continually comb the store to replace products running low on the shelves. Observation and software alerts will take care of that, enabling associates to know what to bring out of the back room before customers show up.
Moving among the customers who'll be absorbing knowledge from the Intelligent Retail Lab's interactive displays, IRL's more than 100 associates will be undertaking these retail experiments every day, getting a firsthand view of what's possible for the future. With technology performing mundane tasks like evaluating if shopping carts need to be corralled, associates will be able to spend more time on tasks humans can do best, such as helping consumers, or adding creative touches to merchandise displays.
According to IRL CEO Mike Hanrahan, "The technology has been built to improve associates' jobs, to make their jobs more interesting, to help them alleviate some of the mundane tasks. AI can enhance their skill-set in a very rapidly changing world."
Taking AI to Walmart Nationwide
In line with its name, the Intelligent Retail Lab is a testing ground and prototype for Walmart's new vision for retail. The exact footprint of the IRL store won't be duplicated across the U.S. Instead, its best findings and developments will be applied and pared down across Walmart's 4,800 existing stores.
Coming back to the meat example, tracking the meat section isn't difficult to scale out to more stores since Walmart only needs a few cameras to do it. Those cameras won't need to be watching all the time either, as they can be put on a loop to turn on every few minutes and grab a snapshot. So, battery-powered cameras could do the job instead of hardwired ones, saving Walmart the trouble of wiring Mount Everest into every store. And a limited number of snapshots in each location reduces the overhead associated with so much data collection.
Walmart's nationwide move toward artificial intelligence will also be largely invisible. When Walmart does roll out systems like the meat tracker to more locations within the year, "It'll look like some cameras pointing at a shelf. Customers probably wouldn't notice," says Hanrahan.
Keeping It Simple and Respecting People's Privacy
So, while it's powered by cutting edge tech, the visible near future of Walmart is just a bunch of cameras pointing at inanimate objects on a shelf. And the AI on offer will be dedicated to simple tasks like monitoring inventory and easing the movement of customers through the store.
In contrast to the Amazon Go approach, Hanrahan insists that Walmart is not doing any sort of human tracking at all. The IRL store specifically doesn't have cameras near restrooms, the pharmacy, or employee break areas, to emphasize privacy.
Regarding any future use of AI or surveillance technology, Hanrahan does confirm that "we're not specifically taking anything off the table." But if Walmart were to eventually perform facial recognition or other analysis on customers, it would "absolutely" ask for consent first. "We could never do anything like that unless customers especially opted into that and it was going to create an experience that was really valuable to them," says Hanrahan.
For now, work at the Intelligent Retail Lab is firmly rooted in using the latest AI technology in a back to basics approach to inventory management, while optimizing the experience of Walmart customers and associates.