The Agentic Review

Enterprise — NOVEMBER 21, 2026

Walmart says Sparky usage rose 129%, with larger baskets

The retailer’s figures put AI-led product discovery on the growth agenda, but do not establish that its shopping assistant caused higher spending.

Walmart CEO John Furner said November 19 that U.S. usage of its Sparky shopping assistant rose 129% year over year, with associated baskets 40% larger, during its fiscal third-quarter earnings call, according to PYMNTS.

Those are company-reported figures, not evidence that the assistant caused higher spending. Walmart did not state a usage base in the earnings-call passage. For small retailers, the disclosure is a reason to investigate AI-assisted product discovery, not a revenue forecast.

The Sparky figures are separate from Walmart’s broader retail measures. Across Walmart U.S., transactions increased 1.5% and average ticket rose 1.4%, excluding fuel, according to PYMNTS’ earnings coverage. Walmart’s announcement reported U.S. eCommerce growth of 23% and marketplace and fulfillment-services net sales growth of 49%, the third consecutive quarter above 45%.

In the corrected earnings-call transcript (FY27 Q3), John Furner, CEO of Walmart, said the company is building an environment where agents help customers discover items and decide what to buy. He said Sparky added voice and image shopping, Spanish and French, and expanded to Mexico, Canada and Sam’s Club in the United States.

Furner also described an Agent Gateway for approved third-party agents, with customer profiles, spending limits and access controls. Shoppers will be able to track agent purchases separately and revoke access, PYMNTS reported. Those controls describe how agent access would be managed; they don’t establish how much additional spending agents might generate.

Delivery capacity is another part of that shopping experience. Walmart U.S. President and CEO David Guggina said its delivery operations reach 96% of U.S. households in three hours or less and 60% in 30 minutes or less, CX Dive reported.

“Whether customers shop directly with us or through an agent, they should be able to count on accurate pricing, reliable availability and clear delivery options,” Furner said, according to CX Dive.

For a small consumer-facing business, that suggests a practical measurement task: check whether products surface in AI-assisted shopping journeys, then measure referred visits, engagement and completed orders separately. Larger associated baskets alone can’t answer whether an assistant brings in new customers or converts visits into purchases.

The disclosure also gives owners a reason to assess merchant readiness for agent commerce. The business question is narrower than whether shopping agents are gaining attention: can customers find the products, see dependable purchase information and complete an order?

Sources

— END —