Walmart's AI Shopping Assistant Sparky Drives 35 Percent Higher Order Values — And Half of Walmart App Users Have Already Tried It

Walmart's AI shopping assistant Sparky has delivered one of the most concrete performance benchmarks in retail AI: customers who use it build baskets roughly 35 percent larger than those who do not, CEO John Furner disclosed on the company's fourth-quarter earnings call on February 19. About half of Walmart's app users have already tried Sparky — making it one of the most widely adopted consumer AI tools in retail, less than a year after its June 2025 launch.

The numbers matter because they are not projections. They come from a public earnings call, verified by more than ten independent retail publications including Modern Retail, Retail Dive, CX Dive, CIO Dive, and Constellation Research. This is measured performance at Walmart scale.

What Sparky does that traditional search does not

Sparky is not a search bar with a chatbot wrapper. It is an agentic shopping assistant — it plans purchases, adds items to your cart, synthesizes product reviews, and adapts recommendations based on what you are trying to accomplish.

David Guggina, President and CEO of Walmart U.S., described the shift on the same earnings call: Sparky is helping Walmart "evolve from traditional search to intent-driven commerce." The difference is structural. Traditional search: you type "birthday candles" and browse results. Intent-driven commerce: you tell Sparky "I am planning a birthday party for twelve eight-year-olds in my backyard" and it builds the full basket — plates, napkins, candles, decorations, snacks, drinks — then offers delivery in under three hours.

Furner put it directly: "Sparky can help understand really clearly what it is that you're trying to accomplish in your life, whether that's a birthday party, or a camping trip." The AI does not just find products. It understands the occasion and assembles the solution.

The logistics back it up. Walmart's express delivery customers grew more than 60 percent year-over-year. Sparky builds the basket. Walmart's delivery network gets it to your door. The AI engagement converts into a physical outcome the same day. Sparky is currently US-only, with Furner indicating plans for global expansion.

2026 is the validation year for AI checkout

Sparky is one front in a broader shift. 2026 is the first full year consumers can complete purchases inside AI platforms themselves — not just discover products, but buy them without leaving the conversation.

ChatGPT's Instant Checkout feature, launched in January, connects to Target, Etsy, Walmart, Instacart, and Shopify merchants including Skims and Glossier. Perplexity opened its AI checkout to all US users in late 2025. Google's Gemini surfaces Walmart products directly. Surveys from the 2025 holiday season found that between 33 and 83 percent of consumers used AI tools in their shopping process, depending on the survey and definition.

But the brand side of this equation is uncomfortable. Brooklinen COO Rachel Levy told Modern Retail: "I want to know who's new and who's a repeat customer." Everlane CFO Vince Adams echoed the concern — brands "want to understand the person behind the click." When a consumer buys Brooklinen sheets through ChatGPT, Brooklinen gets the order but loses the customer data, the browsing behavior, and the direct relationship. The trade-off is real: broader reach through AI platforms at the cost of knowing your own customers.

Does a 35 percent bigger basket serve the consumer?

Sparky's 35 percent lift in order values is good for Walmart. Whether it is equally good for the consumer depends on what is driving that increase.

If Sparky helps a parent assembling a birthday party find items they would have forgotten — the trash bags, the ice, the backup juice boxes — then the higher basket reflects genuine utility. The consumer spent more because they needed more, and Sparky saved them a second trip. If Sparky is optimized to suggest premium alternatives or add-ons that serve Walmart's margin goals, the higher basket serves the retailer more than the shopper.

Walmart's own framing leans toward the first interpretation. Furner described applying AI to "very specific areas where it can make a difference" — a deliberate, use-case-driven approach rather than a blanket AI overlay. That specificity is building adoption: half of app users engaging within nine months suggests Sparky is delivering enough value to earn repeat use.

But the structural question does not go away. When the AI shopping assistant is built, trained, and deployed by the retailer, whose interests does it optimize for when those interests diverge from the consumer's?

Mubboo's Take

Walmart's Sparky at 35 percent higher baskets and Macy's Ask Macy's at 4.75x spending increase tell the same story from two data points: AI shopping assistants built by retailers increase spending. That is their purpose — they are designed to sell more, and they do. An independent comparison platform serves a different function. Not to build a bigger basket, but to help the consumer build the right one. At Mubboo's Shopping channel, our recommendations do not benefit when you spend more. They benefit when you make a decision you return satisfied with — because satisfied consumers come back. The 35 percent higher basket is Walmart's success metric. Consumer satisfaction with the purchase, not the purchase size, is ours.