AI Momentum 107.6 +2.7%GitHub Pulse 5172 -5.9%Capital Flow $204.6B +5804.2%Hiring Heatmap 0.0 -100.0%
Advertisement
Enterprise Adoption

Walmart, Target and Lowe's say their AI shopping assistants lift sales. None has said how much revenue they bring in

Three of America's largest retailers reported quarterly results this week, and all three pointed to their AI shopping assistants as a reason customers are buying more. For investors, these are the latest numbers from companies at this scale on whether chatbot-style shopping pays off.

The figures are striking. They are also hard to judge. Each one describes shoppers who chose to use the assistant. Target and Lowe's compared them with shoppers who did not; Walmart did not spell out its baseline on the call. None of the three said how much revenue flows through its assistant, or what it costs to run.

What each retailer claimed

Walmart reported on Thursday. Revenue reached about $188.8 billion, ahead of analyst forecasts, and the company raised its full-year outlook. Global online sales grew 17%.

Chief executive John Furner told analysts that US use of Sparky, the company's shopping assistant, grew 129% from a year earlier. He said shoppers who use it fill baskets about 40% larger. Sparky has also launched in Mexico, Canada and at Sam's Club in the US. It now accepts voice and photo requests and has added Spanish and French.

The basket figure is not new. Walmart cited the same 40% in August, when it said Sparky users were up 70%. So adoption is speeding up, but the reported lift per order has not changed.

Target reported on Wednesday, with comparable sales, from existing stores and its website, up 5.6%. Chief executive Michael Fiddelke said guests using Ask Target, an assistant still labeled a test version, move to purchase at about twice the rate seen elsewhere. Target did not say how many people use it.

Fiddelke named ChatGPT, Gemini and Copilot as partners for shopping that starts inside other companies' AI apps. He was candid about scale: on that kind of shopping, he said, "volumes aren't high yet."

Target's clearer AI gain may be behind the scenes. Its chief operating officer said software-driven buying and restocking decisions now cover more than 90% of its range of products, up from about 40% a year ago. She said the share of items in stock on shelves rose by more than a percentage point.

Lowe's, also reporting Wednesday, said its Mylow assistant, which it says was launched on OpenAI's platform, answers more than a million customer questions a month. Chief executive Marvin Ellison said online shoppers who use it convert at triple the rate of those who don't. Lowe's online sales rose 17.4%. It gave no user count and no sales figure, and said it would say more at a December investor event.

Home Depot, which reported Tuesday, offered intent rather than numbers. Its incoming chief executive, Richard McPhail, said the company will be "AI-enabled." Other executives cited AI search and an AI tool that builds materials lists for contractors.

Why the numbers prove less than they seem

The problem is who chooses to use these tools. A shopper who opens an assistant to ask about a drill or a holiday gift is often already close to buying. Comparing that shopper with every other visitor mixes the effect of the tool with the habits of the people drawn to it.

The standard fix is a holdout test: randomly hide the assistant from a slice of shoppers and compare the two groups. None of the four retailers said it had done this. Walmart and Amazon each cited 40% order lifts this summer, and enterprise AI writer Rajesh Beri noted in August that neither had disclosed such a test.

Costs are missing too. Running a large AI model for every question is not free, and none of the companies broke out that spending. When an analyst asked Walmart about savings from automation or AI, the finance chief answered with figures on the growing share of online sales and more than $2 billion in extra fuel and transport costs. He put no number on AI savings.

Walmart did credit supply chain automation for lowering the cost of filling online orders. That is a long-running investment, separate from the chatbot.

What buyers and investors should watch

The bigger financial question sits in advertising. Walmart's US ad sales grew 37% in the quarter, including its Vizio TV business. Walmart Connect, which sells ads to shoppers searching and browsing Walmart's site, grew 36%. If customers' own AI agents do the shopping instead, those ad slots could matter less.

Walmart's growth chief said the company is "still learning a ton" about how conversational requests differ from traditional search. Furner described new settings that let customers give their own agents permissions and spending limits on Walmart's site.

Three near-term markers will show whether these assistants are a sales engine or a feature. Lowe's December investor event is the first chance for a revenue number. Target's move out of test mode, and any user count, would show scale. And a retailer publishing a proper holdout result would set the standard others would be pressed to meet.

For companies outside retail weighing similar tools, the lesson cuts both ways. The biggest retailers now treat AI assistants as standard equipment. They have not yet shown the kind of proof a finance chief would ask for before a large budget.