Walmart has announced that it has used large language models (LLMs) to automate and complete important tasks that would previously have required 100 times as many employees to complete manually.
In a recent quarterly earnings call, Walmart executives revealed details of their ambitious automation program, announcing that generative artificial intelligence enabled the supermarket chain to implement a major transformation project while using just 1% of the staff required in the pre-AI era.
The company uses Large Language Models (LLMs) created by third parties and creates its own, for example to improve its product catalog.
Stephanie Wissink, Senior Vice President and Head of Investor Relations, said: “The quality of the data in our catalog impacts nearly everything we do, from helping our customers find and buy products, to how we store inventory across the network, to how we deliver orders. We have used several large language models to accurately create or enhance over 850 million records in the catalog.
“Without the use of generative AI, this work would have required almost a hundred times the current number of employees in the same amount of time.”
Wissink added: “As we look to strengthen our business for the future, we continue to be pleased with the automation work in our supply chain and our overall technology advancements. We are finding concrete ways to use generative AI to improve the experience of customers, members and partners.”
By year’s end, 3,000 of the company’s 4,600 stores will receive their deliveries from automated systems, said John Furner, president and CEO, who also highlighted GenAI’s productivity gains.
He added: “Using GenAI is about 100 times more productive than having people go through every single product display page… The GenAI product we used helped us capture the attributes and features of hundreds of millions of items. That would have taken, as I said, 100 times longer if we had tried to do that manually.”
John David Rainey, executive vice president and chief financial officer, also said Walmart is “making good progress in transforming our supply chain.”
In the U.S., over 45% of e-commerce fulfillment center volume is now automated, and about 1,800 stores “receive some level of freight” from 15 regional distribution centers that “are in varying stages of automation implementation.”
“And while we are spending more on capital expenditures than ever before, we are pleased with the return on those investments, particularly in automating our supply chain,” Rainey continued.
Walmart reported second-quarter revenue of $169.30 billion, beating analysts’ forecast of $168.56 billion.
In June 2024, the chain announced that it was developing a shopping assistant using GenAI technology that would allow customers to ask questions like “Which TV is best for watching football?” and then be shown a TV that meets their needs. The chain has also introduced GenAI-based search and AI product reviews, summaries and comparisons.
All of Walmart’s GenAI tools are built on proprietary AI/ML and conversational AI platforms.
In a letter to shareholders released alongside its earnings report earlier this year, Walmart committed to investing in the technology sector, saying the company “employs thousands of people in technology roles around the world.”
“They innovate to bring new experiences to life, gain insights from data to enable our decisions, and intelligently automate our operations with technologies like generative AI, computer vision, and automated warehouse and retrieval systems in our distribution and fulfillment centers,” wrote CEO Doug McMillon.
“We have introduced innovative features such as generative, AI-driven product search, which enables customers to shop more intuitively.”
He added: “As we advance our digital business, we are redesigning the business model. Think of it as a combination of a traditional retail P&L paired with a newer version that starts with our digital business.”
