Walmart’s Intelligent Retail Lab store runs on AI
A Walmart Neighborhood Market has opened in Levittown, NY to not only test artificial intelligence (AI) in physical settings but witness how such technologies impact the customer and associate experience.
Called the Intelligent Retail Lab (“IRL” for short), the store looks like an Amazon Go on a broader scale, covering over 30,000 items across 50,000 square feet. According to a Walmart release, the sensors, cameras and processors are connected “by enough cabling to scale Mt. Everest five times and enough processing power to download three years’ worth of music (27,000 hours) each second.”
However, the store has manned registers. Instead of supporting cashierless checkout, the technology will initially monitor inventory, shopping carts shortages and cash register lines.
Walmart’s plan is to use the cameras and real-time analytics to automatically trigger out-of-stock notifications to internal apps that alert associates when to restock. The technology will need to recognize the specific product (i.e., distinguishing between one pound or two pounds of ground beef) and compare the quantities on the shelf to future sales demand.
“A Glimpse Into the Future” – Source: Walmart
That’s expected to spare associates from walking the aisles looking for shortages. Other applications being tested are designed to make sure shopping carts are available and registers are open.
Walmart also noted that an AI-driven shopping environment “also raises questions about all the visible technology,” which is intentionally in full view. The store includes a “glass-encased data center bathed in blue glow,” educational kiosks throughout the store, and a Welcome Center up front for a deeper dive into the technology. An interactive wall prompts customers move around to see how AI tracks their body positioning.
Finally, Walmart will see how the more than 100 associates at the store interact with the technology that’s expected to reduce their mundane tasks.
Walmart cautioned that IRL’s focus will be on data-gathering in the near term. Mike Hanrahan, CEO of IRL, said, “There are a lot of shiny objects out there that are doing things we think are unrealistic to scale and probably, long-term, not beneficial for the consumer.”
- Walmart’s New Intelligent Retail Lab Shows a Glimpse into the Future of Retail – Walmart
- Walmart unveils an AI-powered store of the future, now open to the public – TechCrunch
- At Walmart’s Intelligent Retail Lab, the future is here – Supermarket News
- Walmart Rolls Out Futuristic AI-Powered Concept Store, Hoping to Best Amazon Go – Inc.
Discussion Questions
DISCUSSION QUESTIONS: What do you think of AI’s potential to monitor in-store out-of-stocks, shopping carts shortages and cash register lines? Is making the technology visible to shoppers a smart move, and should tech be more out in the open at regular stores?
Poll
BrainTrust
Bethany Allee
Senior Vice President Marketing, PDI
James Tenser
Retail Tech Marketing Strategist | B2B Expert Storytelling™ Guru | President, VSN Media LLC
Walmart’s investment in learning and innovation is proving again why it’s a leader – this test store is a good example. Lots of retailers have lab stores, however the level of commitment that Walmart exhibits and their approach to displaying the technology in action is what makes it different.
While there’s plenty of speculation about how AI can be applied in retailing to eliminate mundane tasks and understand how shoppers and associates interact with the store, Walmart is actually testing it in a meaningful, open way. I think making the technology visible to shoppers is the right move as it normalizes the technology for shoppers, making it all seem more normal and natural.
There are many intriguing elements to the IRL store including the visibility of the technology (which I predict will be a short-lived novelty in the evolution of AI in retail).
But to the question posed, AI’s potential to monitor OOS, checkout lines, etc. is not in question — retailers ability to manage humans to correct those situations is.
As usual, a story. I was working at Frito Lay when the first handheld computer for route salespeople was rolled out. It was supposed to generate enough extra time in the day for the average route to add about two stops — more than enough to pay for the capital investment. Instead the “start-stop” times recorded by the computers showed that the technology saved enough time in the day for the average route person to return to the DC 30 minutes sooner than before — which is what they did. Home in time for Oprah — yea!
Great observation about the visibility of sensing tech in this lab store, Ben. At this moment it’s very important for shoppers to feel there are “no secrets” about how they are being observed interacting with the merchandise. The wow factor will be exciting to some, too.
This is a great application of AI’s power. The learning Walmart will get from this will be phenomenal and a potential game changer. Certainly learning about the true cost of out of stocks, shortages and time in line will transcend all Walmart stores.
Walmart will learn quickly how the customer reacts to this technology play from a consumer perspective — whether they warm to it or not. Without having the facts, it is mere speculation. But I can imagine that as long as the store is clean, well stocked and the prices are right, the customer will be satisfied.
AI can definitely do the things Walmart is testing. Walmart is brilliant to have technology visible (the data center component) in a store that’s branded as an innovation center. Companies struggle to test and efficiently deploy technology. This store is an extremely sound idea to test the technology, learn, tweak, and then scale. With this approach, Walmart will bypass or minimize the impact of miss-steps. AND there will be AI miss-steps.
I’m such a nerd, I want to go and see this in action.
I applaud Walmart for taking this on. Not unlike last week’s “AI-powered, voice-capable chatbot helps shoppers make the right choice in stores” discussion, this too is the future of retailing. Not the near future. This is a lab funded by Walmart and thus can support the expense and effort to determine the value of such a store, the challenges to run/scale it, and the costs.
A decade or so from now, no one will be disputing the prevalence of AI anymore (probably still the invasiveness though) and the thought of most retailers having some form of “intelligent stores” will be as mundane as most people having mobile devices. From now until then, there will be a long and treacherous journey and the end result (there never really is an end) won’t be anything like this prototype.
So for the naysayers, you can argue against it, you can dismiss it, but the genie is out of the bottle and it will never go back in, no matter how poetic the old days of human-only stores seem.
For customer facing engagement, these are the places to start — out-of-stocks, lines, and physical carts. Why? Because that’s what the customer engages with. They don’t see the behind-the-scenes operations, logistics, or employee productivity parts.
Transparency by Walmart is a great move and differentiates from closed systems in-store where all the customers know about is a black box system that powers the store convenience (e.g. Amazon Go). Customers already know tech is there, making it open fosters trust and reduces concerns about privacy. As customers understand the data being collected is to better their in-store experience, it becomes a non-concern for them. Great lab setup for Walmart and will provide them with a good perspective for focusing on the right store tech.
Artificial Intelligence is becoming the new Omnichannel trend of the moment in the retail world. After so many years of hype, consumers are finally reaping the benefits of a Omnichannel-ready retailer, with BOPIS and other fulfillment options.
As they go through their maturity cycles, the same could be applied to AI, AR, and any other digital innovations retailers are looking to integrate into their operations. To answer the question, the biggest benefits of AI are the predictive analytics that retailers could operationalize around to provide a better customer experience.
Purpose led AI, solving an actual business challenge is what matters. While Walmart has opened the hood to show the technologies to the customer, it’s not necessary to be so transparent. Consumers still want the right products, at the right price, and at the place they wish to engage and transact.
One could write a book to answer the question. AI can have a big impact on reducing out-of-stocks, identifying shopping-cart shortages and the wait at (and activation of) cash register lines. The fact is that the execution of the response to all the triggers as well as the high cost of implementing these enhancements will be the main factors to their success, or failure.
If their version of AI can resolve issues like products in the wrong place, not in a straight line, not facing perfectly front, etc. then this could be a very cool application. The question then is whether it is less expensive than a manager walking around a store making notes or staff with assignments that would be regularly checking. You still (today) need someone to get it out of the backroom and load the shelf, or someone to place an order.
There are few retailers with the resources to devote to such an expensive experiment, but Walmart shows us why they are a retail leader with their IRL store. Walmart will no doubt gain tremendous knowledge from the data gathered from so many AI-based approaches in the store. The real challenge will be how they then turn that data around into insights they can apply across their entire fleet of stores.
This is a human processing issue, not a computing power one. If they succeed, there they will gain significant advantages over competitors in their ability to quickly respond to customer needs in the store, especially around out-of-stocks. I also expect Walmart to learn quite a bit about their customers’ path through the store and how the store layout and merchandising affect this path. The key here will be how the AI insights are translated into actionable changes that store staff can implement. Whether there is a real-time need for this is still an open question for retailers and Walmart may gain important data that helps them understand what the use of such real-time knowledge would do for them.
Making the technology visible to customers is an interesting novelty that will surely attract more customers than it will detract. It also allows for some good PR about what and how Walmart is using AI that will make customers feel more comfortable about the technology and its use. I do not see this as a trend however, but more as a novelty for this specific application.
AI makes a ton of sense for monitoring out-of-stocks and other in-store issues. What Walmart is doing is a smart move, and its lab store is a good way to ease consumers into the idea of visible AI. Overall, more retailers should have their technology front and center, as long as it isn’t literally in the way. Consumers will learn to live with the technology and retailers will have more options to develop and test technology as consumers get comfortable.
I absolutely love this. It is an amazing example of how technology can be used to remove menial tasks. It means staff can spend time on customer service. It will need upskilling of staff — but this is the same in many industries and skills like emotional intelligence become more critical.
This test/lab store seems much more operations oriented than customer need driven. Granted, out of stocks can turn a customer off, but not having a much swifter checkout system (a la GO) as an example, is not a tip of the hat to the most important person in their lives. And frankly, the big “show” of all the tech in the back is, self-important vs solving any customer issue. Let’s hope that their future efforts at this make the right corrections.
I think it’s important to put examples like this into context. Walmart’s investment in analytics and AI is probably only matched by Amazon, to a lesser degree Kroger. These use cases are terrific, but require a LOE that few retailers or their CG partners can understand, let alone have the resources to execute.
Most companies need to step back and check their analytics maturity while mapping out a plan to leverage data inline with their business strategy. The first use cases that fall out of this exercise will focus less on complete store transformation like this Walmart example, and more on pilots for marketing personalization or supply chain planning. Success with initial use cases funds the type of lab concept you see here.
Walmart’s application of AI relative to Amazon is notable by retaining the use of store associates as part of the overall CX. It’s also helpful to showcase the technology and tracking in the store itself to overcome customer concerns about transparency and trust.
What does a good store manager know? Traffic by day part, by day or the week. Demographic profiles — again by day part, by day of the week. Operational issues — on-going — baskets, lines at checkout, chronic out of stocks. Generally predictable with 70% accuracy to a good store manager. What does getting to a high percentage do? Not much. Retail’s/AI’s challenge is to know what the manager doesn’t know that is actionable with available resources. We are looking for a better mix of Art and Science. We need to start with what we already know ….
I’m stuck in Arizona with the Long Island blues again, wishing I could run right over to Levittown and give this place a close personal inspection.
There’s so much to say about Walmart’s IRL experiment. Number-one for me is the commitment the company is making to in-store sensing, which I believe is an area of innovation whose time has come.
The ability to monitor inventory on display is a crucial discipline that all retailers must aspire to. I’m especially intrigued to learn whether IRL can provide actionable signals about perishable product freshness and display replenishment.
I grow weary, however, about the pedestrian focus on out-of-stock detection. By now, the industry should understand that detecting OOS is like shutting the barn door after the horses are out. The real opportunity lies in managing on-shelf availability with real-time perpetual store inventory and automated reordering. OOS prevention must be the primary objective, which will make detection and remediation rare events.
Yes I know I tend to go on about this topic, but if the data flows from this IRL experiment are as comprehensive as promised, real time PI may finally become practical at scale. Unprecedented inventory accuracy and service levels will surely follow.
AI has the potential to optimize unlimited tasks and make more accurate decisions than a human, or a team of humans, can. The technology for AI tools has advanced dramatically in the past five years: however, it is still underutilized by most retailers.
According to BRP’s 2018 Integrated Planning and Inventory Management Survey, most retailers (67%) are not leveraging advanced analytics to improve their planning decisions and optimize inventory and only 39% of retailers identified improved analytics as a top priority. This is a disconnect. As technological capabilities continue to advance, investing more resources into data utilization needs to be a critical objective for retailers.
Walmart is very methodical with their innovation testing and making the technology visible in the IRL is another test. Measuring the impact of many options enables companies like Walmart to test many concepts until they find the best ideas. It would be interesting to hear what feedback they get from the customers on the visible technology. For some customers, I suspect they will perceive Walmart as more innovative and progressive. There is never a dull moment in retail.
The need is for real intelligence as in managing out-of-stocks, customer service and checkout time and ease.