Walmart executives seem pleased with the results of using generative AI to rewrite product listings, saying the task is much faster than human labor. They also plan to launch a new shopping assistant based on the technology.
Walmart President and CEO Doug McMillon told investors during the company’s fiscal second-quarter earnings call on Aug. 15 that the company used several large language models to create or enhance more than 850 million records in its product catalog.
“We are finding concrete ways to use generative AI to improve the experience of customers, members and partners. We are leveraging data and large language models from others and building our own,” said McMillon. “Without the use of generative AI, this work would have taken nearly 100 times the current headcount in the same amount of time.”
John Furner, president and CEO of Walmart US, said generative AI helped the company capture the attributes and characteristics of hundreds of millions of items, and doing it manually would have taken hundreds of times longer.
Furner says this initiative has helped the company better understand what customers want on the website and in-store with better product pages. “We can tailor the catalog to their needs much more effectively because the details of each item and the product display pages have become much better.”
Retailers have quickly adopted generative AI, popularized by ChatGPT in November 2022, to provide more personalized search tools and virtual assistants online and generate marketing content. Target, Walmart and Best Buy, as well as smaller retailers like Boot Barn and Tractor Supply Co., have also announced plans to introduce AI-powered tools to improve the shopping experience, especially for employees. Amazon uses AI to rewrite product listings that don’t meet its requirements.
“The use cases for this technology are diverse and affect nearly every area of our business. We will continue to experiment with and deploy AI and generative AI applications around the world,” said McMillon. “We are committed to using AI responsibly, but at the same time we are moving quickly and (in a practical, cost-effective way) to meet our future needs and scale these experiences.”
Walmart already has an AI-powered search function in its app and on its website, and promises that a new shopping assistant can help with questions like “Which TV is best for watching sports?” and ask follow-up questions like “What is the lighting like in the room where you’re putting the TV?”
McMillon said the company is also testing a new experience for sellers on its marketplace where they can ask the company questions through a new assistant that provides concise answers they would otherwise have to search for in lengthy articles or other materials. “We want our sellers to focus on selling, so the more seamless we can make it, the better,” McMillon said.
Walmart’s e-commerce business grew 22% globally from the second quarter of 2023 to 2024. Furner said the use of generative AI for its product catalog has been a “huge enabler” for the e-commerce business in recent months.
The productivity gains to be achieved through generative AI are “really important in the context of the market, where we have greatly expanded both the number of sellers and the number of items in the assortment,” Furner said. “We increasingly feel that Walmart can sell our customers anything they are looking for.”
Bryan Gildenberg, founder of commerce consultancy Confluence Commerce, said Walmart, like other companies, has found that the most interesting applications of generative AI are commercial or B2B-focused rather than consumer-focused. He said applications such as helping people decide which TV to buy or Amazon’s conversational assistant Rufus are difficult for AI to master quickly.
Gildenberg said that like humans, AI is most effective when you give it a clearly defined task, teach it what to do, have clear expectations and have patience to learn. “AI is really good at learning a specific task, and if you focus AI on a very specific task, it learns really quickly,” he said, adding that Walmart’s catalog workflow automation – a simpler task – “is the kind of application where AI can be a game-changer.”
Gina Logan, senior retail analyst at Kantar, said she finds it interesting to see Walmart relying so heavily on AI while McDonald’s recently decided to back away from AI ordering technology it had been working on with IBM. McDonald’s did not disclose a reason for doing so, but it came after some ordering errors with the AI system went viral online. In one popular TikTok, for example, the AI constantly confused an order of “Mountain Dew” with “Medium Coke.”
“It sounds like Walmart is confident in the implementation and really wants to leverage it in a way that they can refine their business strategy, particularly their marketplace,” Logan said. “I think that’s an investment they’re making to improve the employee experience, and I think that will hopefully translate into better customer service for them.”
Aside from AI, the retail giant also announced in its second-quarter earnings statement that its revenue increased nearly 5% year-on-year and forecast an optimistic but cautious view of consumer demand.
“We see our members and customers remain selective, discerning and appreciative, focusing on things like essentials rather than unnecessary items. But importantly, we are not seeing any additional impact on consumers’ health,” Walmart Chief Financial Officer John David Rainey told CNBC.
