AI is embedded everywhere at Walmart


At Walmart, artificial intelligence (AI) and machine learning are everywhere.

You won’t see it walking down the aisles of a Walmart store. You don’t feel it when you pick up a Walmart package from your doorstep. And you won’t notice if you search Walmart’s website for everything from paper towels to toys.

But today, AI and ML are embedded throughout the Walmart organization — from supply chain management and shopping to search.

as the the world’s largest retailerIt’s no surprise that the Bentonville, Arkansas-based retail leader has been investing in advanced AI for years: In 2017, for example, VentureBeat highlighted Walmart’s massive inventory boost thanks to AI, while reviewing Walmart’s AI efforts on everything from express delivery to grocery delivery robots in the past half decade.

And in the past six years, Walmart has gone from a handful of in-house data scientists to hundreds, according to Srini Venkatesan EVP, US Omni Tech at Walmart. These data scientists work in teams dealing with supply chain forecasting, optimization, and labour/demand planning; search and personalize; as well as emerging technologies. “We really spend a lot on internal development because we think this is our competitive secret sauce,” he told VentureBeat.

Venkatesan, who leads all of the technology teams powering Walmart’s global marketplace, omni-supply chain, and stores, said Walmart is “evolving from a retail automator to a retail enabler — that’s where AI and ML are very relevant to us. What he means by ‘automate’ versus ‘enable’, he added, is that the company has moved from simply using technology to make Walmart’s tools and processes more efficient, to stepping back.” We’re looking at the overall end-to-end picture to enable improvements across the store journey and across the organization,” he said.

Which, of course, leads directly to Walmart’s biggest goal: figuring out what the customer wants and offering it. “Walmart has always been about what the customer wants,” he said. “The customer is number one.”

AI permeates Walmart’s entire supply chain

To give customers what they want in an era of global supply chain problems, Walmart has put an emphasis on AI in the supply chain: last week the company announced that it four next-generation fulfillment centers (FCs) over the next three years, with its first debut this summer in Joliet, Illinois.

These FCs will be the first of their kind for Walmart, using robotics and machine learning to accelerate execution. The company claims that Walmart, combined with its traditional fulfillment centers, can now reach 95% of the US population with shipping within the next or two days.

In addition, Walmart announced last Monday that it will bring Symbotic’s next-generation robotics and AI technology to all 42 regional distribution centers — it’s already been in use in 20 — over the next eight years as the retailer works to modernize its supply chain network. . The technology should help Walmart increase its inventory accuracy and increase the capacity of its warehouses to receive and ship products to stores, the company said in a statement.

symbolic, which became public this week and enjoy a significant investment of Walmart, said its AI-powered software and robotic system — including its Symbots, fully autonomous vehicles that leverage machine learning, vision and algorithms — addresses some of the biggest challenges of Walmart’s complex supply chain.

“If you look at things like accuracy and fewer errors and less downtime, there’s just incredible savings from a working capital, inventory management and total labor perspective,” said Symbotic CEO Michael Loparco. “So I think there are powerful cost drivers, but I think the biggest catalyst for Walmart is changing consumer demand and the need to push the market.”

Walmart’s Evolved Supply Chain

Walmart’s supply chain efforts using AI have evolved in recent years, Venkatesan said, from simply forecasting sales demand — how much will be sold that’s already in stores — to predicting consumer demand in terms of what the customer actually wants to buy, by analyzing data across various channels, from Google searches to Tik Tok social feeds.

During the pandemic, however, a difficult demand problem to solve also became a thorny supply problem.

“We learned that we needed to understand what would be out of stock and what to replace it with,” he said. “So we invested a lot of energy in AI and ML for replacement logic.” Deep learning AI considers hundreds of variables – including size, type, brand, price, customer data collected, individual preferences and current inventory – in real time to determine the best next available item.

AI powers Walmart’s search and personalization

Historically, much of Walmart’s search and personalization activity has revolved around automated decision-making, said Jan Pedersen, VP of search and personalization, US Omni Tech, at Walmart. But more recently, the performance of computer vision AI models has become much better than it used to be because of deep learning, he explained. “You can use these things in production and get results,” he said.

As a result, there are several areas where Walmart is using AI technologies and natural language processing in search and personalization, he explained. There are English searches – understand what people mean when they request a product type, understand which parts of the search are important.

Understanding the quality of the image is also key, he added. “Maybe even attribute extraction, so it’s important to know it’s a red shirt because it’s red in the picture.” Finally, there is the machine translation. “We don’t have to translate anything manually, so that’s a big boost,” he said.

Search is a growing frontier

However, some questions are much easier than others, he stressed. “Maybe you have a question that is often repeated and people are giving you a really strong signal about what it means, or you might have a question where, if you look at it, the intent is very clear what the user is interested in, but if you if you attack it from a standard approach, you don’t really get good results.”

An example of this, he recently explained, was “avocados from Mexico.” “The reason that’s interesting is that most avocados don’t tell you they’re from Mexico.” On the other hand, he explained, the question itself is very clear: it is clear what the user wants. “So we put that in the bucket of semantic questions where you really need to be on top of it, understand the avocado part is important or generally infer from other things you know about an item that’s probably important.”

Finally, Pederson discussed Walmart’s efforts toward multilingual search, allowing Hispanic customers to find specific items on the Walmart.com site and app.

“One of the interesting things about search experiences in general is that people can type whatever they want because it’s an empty box,” he said. To serve customers seeking Spanish, Walmart uses language detection using AI. “You find out that this search is probably in Spanish and then you translate the search into English,” he said. “Then, when we have the results, we return the results in English. The next level is to machine translate the content of the product description so that we can translate the titles.”

AI-powered fitting room technology

Computer vision also powers one of Walmart’s most recent AI-powered offerings: Zeekit’s dynamic virtual fitting room technology, which Walmart acquired last year. It allows customers to buy clothes online and see what an item actually looks like.

Walmart’s “Choose My Model” experience, launched in March, offers customers a choice of 50 models ranging in height from 5’2″ – 6’0″ and sizes XS – XXXL. Customers can choose the model that best suits their height, body shape and skin tone.

“Based on the millions of images we have from the catalog, we analyze all of the different articulation points on the models and use those to create the dress simulation,” said Desiree Gosby, VP, new businesses and emerging technology at Walmart. “It’s about breaking down everything from whether it should fit loosely or not, where the waist should fit, how the length should be adjusted depending on your height.”

Gosby’s team is currently working on an experience using Zeekit’s technology where customers can upload their own photos. “It’s actually a more difficult problem for AI and computer vision,” she said. “And customers need to make sure they get a good shot that makes them feel good.”

Conversational AI in the mix

Walmart also recently launched its conversational AI technology called Text to Shop after several months of testing. Customers can text or say what they need and Text to Shop will add it to their cart. If they need an item they’ve never bought before, Walmart provides product recommendations

“This is really about how we make it easier for customers to express what they want or need from us,” said Gosby. “It’s basically a digital assistant platform that uses voice and text chat — we work with the entire company, including customer service, and we power Walmart’s shopping assistants on Google and Siri.”

Text to Shop is the result of a lot of investment in understanding natural language, she added. “We use GPT-3 under the hood and then really leverage our data to create natural language comprehension that is natural.”

But, she admits, “it’s actually really hard to make this simple — being able to understand when you say things like chocolate milk and add pizza to my cart that you really mean chocolate milk as opposed to chocolate versus milk.”

In general, these technologies are about giving customers confidence to make purchases, Gosby said. “Do we really save them time? Do we lower the return rates for clothing?” Everything Walmart does should be aimed in some way at removing friction for the customer, she said, “We don’t do technology for the sake of technology.”

Walmart’s AI is highly customer-centric

When asked about the future of AI at Walmart, Venkatesan returned to focus on the customer. “Our prediction of the future has always been what the customer wants – we observe the customer very carefully,” he said. “We can understand how customer trends are going and we’ll adapt to that because it’s very difficult to predict exactly where it’s going to go.”

Walmart will continue to refine, he added. “I think there are many more improvements to be made,” he said. “It will be a constant evolution or upgrade of what we do continuously as it only gets more complex as customer requirements change.”

VentureBeat’s mission is to be a digital city square for tech decision makers to learn about transformative business technology and transactions. Learn more about membership.

Leave a Reply

Your email address will not be published.