At a glance
- Company
- WalmartRetailer
- Industry
- E-commerce and retail
- Where and when
- United States. Reported August 2024.
- Who produced the figures
- The company itselfThe figures come from prepared remarks by Walmart’s chief executive on an earnings call with investors.
What they did
A retailer’s catalogue holds the details of every product: size, colour, material, images. Walmart’s chief executive told investors that the quality of this data “affects nearly everything we do”, from helping customers find products to sorting stock and delivering orders.
Walmart used several large language models to fill in and correct this data.
The numbers
- More than 850 million pieces of catalogue data created or improved.
- “Nearly 100 times the current head count”: what Walmart said the same work would have needed, in the same time, without generative AI.
Walmart gave no error rate and did not say how the output was checked.
What happened next
On the same call, Walmart said customers were already using AI-powered search in its app and on its site, and that it was adding a shopping assistant.
What the case shows
- A good first target. The task is frequent, well defined and has a clear standard for a correct result. That matches the advice to start with one costly process.
- AI made work possible that was not being done. Walmart did not say it replaced a team. It said the job was too large to do by hand. No cost was removed.
What this says about the practices
Practices the company applied
Sources
- Walmart Inc. (15 August 2024). Q2 fiscal year 2025 earnings call. Corrected transcript, published on Walmart’s investor site. stock.walmart.com