When the Algorithm Builds the Basket: What Walmart’s Sparky Means for the Suppliers Who Fill It

Walmart’s Q4 fiscal 2026 earnings call, held on February 19, left little room for ambiguity about where the company’s AI shopping agent stands: Sparky is no longer a side experiment. CEO John Furner told analysts that roughly half of the retailer’s app users have engaged with Sparky, and those who do carry an average order value about 35% higher than customers who do not use the tool. Those numbers landed in a quarter where U.S. e-commerce sales grew 27% year over year, according to Walmart’s earnings release. Together, they point to a structural shift in how discovery, basket building, and conversion happen inside Walmart’s ecosystem.

Who these customers are matters just as much as how many of them there are. Furner noted on the same call that the majority of Walmart’s share gains continue to come from households earning more than $100,000 per year. Neil Saunders, managing director at GlobalData Retail, noted in comments reported by Retail Dive that many higher-income shoppers interact with Walmart primarily online, and that this dynamic helped underpin the quarter’s e-commerce growth. Sparky, as a tool embedded in the app, naturally reaches these same digitally engaged customers. The 35% order value premium likely reflects both the agent’s ability to assemble larger baskets and the spending power of the people using it. Brands selling into premium or trade-up categories at Walmart should take note, but the implications run well beyond any single product tier.

Sparky is not a customer experience story. It is an operating model story, and it touches how brands plan, execute, and invest across nearly every part of the Walmart relationship.

The Discovery Layer Has Moved

The playbook for visibility on Walmart.com and the Walmart app has long been anchored in search: keyword optimization, paid placement through Walmart Connect Sponsored Search, and content that performs well against the retailer’s product listing algorithms. That playbook still matters. But Sparky introduces a parallel layer of discovery that operates on different signals.

When a customer describes a need in natural language, asking for help with a birthday party, a week of meals, or a camping trip, the agent interprets that intent and recommends a set of products. It builds a basket. Furner described the sequence on the earnings call: Walmart then executes that basket through its delivery, pickup, or in-store fulfillment network. The process moves from conversation to commerce without requiring the customer to browse, compare, or search in the traditional sense.

This reframes what product visibility means on Walmart’s platform. The signals that inform an AI recommendation are not identical to the signals that drive keyword search rankings. Product attributes, reviews, specifications, and structured data all feed the algorithm’s understanding of whether a given item solves a customer’s stated problem. Brands whose product data is thin, inconsistent, or poorly structured may find their items excluded from AI-generated recommendations entirely, regardless of how well those items rank in conventional search.

Walmart has not published detailed documentation on how Sparky selects products for its recommendations. That opacity is itself a signal worth taking seriously, because it means the selection criteria are likely to evolve without much advance notice, and the bar for content completeness will only rise alongside adoption.

Advertising Inside the Conversation

Walmart Connect posted its strongest quarterly growth in years during Q4. U.S. Walmart Connect revenue grew 41% in the quarter, according to CFO John David Rainey, who described it on the call as the best year-over-year performance in at least three years. Globally, Walmart’s advertising business reached nearly $6.4 billion for fiscal year 2026, a 46% increase reported in the company’s earnings release.

That growth came even as Walmart began testing a format that could reshape how Sponsored Search works: ads inside Sparky itself. Walmart Connect announced in January 2026 that it was actively testing sponsored prompts within the AI agent, where a brand’s product could surface as a recommendation when a shopper asks Sparky for guidance. According to a blog post on the Walmart Connect site that month, 81% of surveyed customers had used Sparky to check product availability or review specifications before buying.

Sponsored Search campaigns optimized for traditional keyword queries may not translate cleanly into a conversational environment, where the shopper’s input is a full sentence describing a need rather than a two-word search term. Walmart has started rolling out an AI advertising assistant, currently in beta for Sponsored Search campaigns, through its Marty super agent for sellers and advertisers. The same Walmart Connect blog post reported that 97% of user queries to the advertising assistant have been unique, suggesting brands are already applying the tool for account-specific optimization rather than generic campaign management.

Furner acknowledged on the earnings call that the industry is still learning how agentic commerce intersects with advertising monetization. But the contours are taking shape. Khurrum Malik, VP of business and product marketing for Walmart Connect, told Adweek in January that the team views Sparky as “an interesting new surface” worth testing into even at this early stage. Brands that begin experimenting with these formats now will build a working understanding of what converts in conversational commerce, which is knowledge that will be difficult to acquire later by reading someone else’s case study.

Your Inventory Is Now Part of the Recommendation

The 35% higher average order value among Sparky users does not exist in a vacuum. Those orders flow into Walmart’s fulfillment network, which is getting measurably faster. Furner reported that the number of customers choosing express delivery, orders arriving within three hours, grew more than 60% in 2025. Rainey added that 35% of store-fulfilled online orders in the U.S. were delivered in under three hours during Q4.

The connection between AI-driven basket building and fast fulfillment is direct. When Sparky recommends a product, Walmart’s system draws on inventory positioned across its network of more than 4,500 U.S. stores and fulfillment centers. If a product is not in stock at the right location, it may not appear in a recommendation for a customer expecting fast delivery. Availability and inventory positioning have always been important at Walmart. They become considerably more consequential in an environment where the algorithm factors deliverability into what it recommends.

Walmart is also building tools to measure and close gaps in physical execution. The same week as the earnings call, Walmart Data Ventures launched Scintilla In-Store, a platform that gives supplier field representatives real-time store-level inventory data, metrics, and task management in a single app. The platform, formerly known as Volt, extends Walmart’s first-party Scintilla insights ecosystem into physical stores. Pamela Stewart, North America Chief Customer Officer for Retail at The Coca-Cola Company, said in Walmart’s announcement that the tool provides “real-time inventory visibility” and supports “data-driven decisions during every store visit.” Future updates will include AI-driven task prioritization. Products promoted through Sparky or Walmart Connect but absent from shelves create a failure that Walmart can now see and quantify, and that visibility changes the accountability dynamic.

Rainey noted that about half of the company’s U.S. e-commerce fulfillment center volume is now automated, and 23 out of 42 regional distribution centers are in various stages of automation. He characterized capital spending on supply chain automation as nearing its peak. The infrastructure that connects digital demand to physical execution is being built at scale, and compliance with On-Time In-Full expectations and inventory accuracy standards feeds directly into whether a product qualifies for the fastest-growing parts of Walmart’s business.

The Agent Is Not Staying Inside the App

Sparky is Walmart’s owned agent, but the company’s agentic commerce ambitions extend well beyond a single app. In January 2026, Walmart and Google announced a partnership at the National Retail Federation’s Big Show that will allow Gemini users to discover and purchase Walmart and Sam’s Club products directly within Google’s AI assistant. The integration uses the Universal Commerce Protocol, an open standard co-developed with Shopify, Etsy, Wayfair, Target, and other retailers. Walmart had already established a similar arrangement with OpenAI’s ChatGPT in October 2025, enabling purchases through that platform as well.

In practical terms, the surface area where Walmart inventory is discoverable through AI is expanding. A shopper who asks Google Gemini about camping gear may receive a Walmart product recommendation without ever opening the Walmart app. The content, product data, pricing, and availability signals that determine which products appear in those recommendations are the same ones that feed Sparky.

That makes this a multi-surface problem, not a single-channel one. Product discoverability now spans a network of AI agents, each with its own interface but drawing from a shared pool of Walmart catalog data. The investment in getting product data right for Sparky pays dividends across every AI surface that connects to Walmart’s systems, and the cost of getting it wrong multiplies across those same surfaces.

The Trust Question Worth Asking

Consumer skepticism toward AI shopping tools is real, and brands should keep it in perspective rather than ignore it. A YouGov survey cited by CX Dive in October 2025 found that two in five American consumers reported having no trust in AI shopping assistants. A Talkdesk survey of 1,000 U.S. consumers, conducted in December 2025, found that 40% would feel misled if a brand did not disclose they were interacting with an AI agent.

Walmart appears to be managing this friction by anchoring Sparky in practical utility. Saunders of GlobalData Retail told Retail Dive that Walmart has addressed skepticism by “applying AI to very specific areas” and solving a real problem with product search and discovery. The engagement numbers and order value data suggest that Sparky has cleared the trust threshold for a meaningful share of Walmart’s digital audience, even if broad consumer ambivalence persists.

The more important question for brands is not whether shoppers trust Sparky. It is whether Sparky’s recommendations will prove reliable enough to shape purchasing patterns at scale. If they do, the brands Sparky recommends will gain share, and the brands it passes over will lose it, gradually at first and then faster as adoption grows. How quickly consumers embrace agentic commerce will determine the size of the opportunity; what brands do now to prepare will determine who captures it.

What This Adds Up To

Rainey was direct on the earnings call about where Walmart’s profit growth is coming from. Advertising and membership fees now represent nearly a third of operating income, and e-commerce has moved past breakeven into profitability with what Rainey described as double-digit incremental margins. The FY2027 guidance projects net sales growth of 3.5% to 4.5% in constant currency, with adjusted operating income growing 6% to 8%, both per Walmart’s earnings release. That spread between top-line growth and profit growth reflects a company extracting more margin from every dollar of revenue, with a significant share of that extraction flowing through the digital and data systems that Sparky, Walmart Connect, and Scintilla represent.

The brands that invest in product content built for AI-driven discovery, that test into Sparky and Walmart Connect advertising formats before they become crowded, that maintain the inventory discipline required for express fulfillment, and that treat their Walmart data infrastructure as a competitive input will be positioned to capture disproportionate share as agentic commerce grows. The ones that wait risk something worse than falling behind: they risk becoming harder for the algorithm to find at all.

Winning With Walmart

Winning With Walmart is an independent platform for suppliers, sellers, solution providers, and industry experts.

Articles are developed by staff editors, contributing experts, and trusted partners.

Some are researched and drafted with the assistance of AI tools and are reviewed, fact-checked, and edited by Winning With Walmart editors before publication.

Editorial judgment, sourcing decisions, and final approval rest with the publication's human editors in every case.

We are committed to accuracy and fairness. If you believe this article contains an error, we welcome your feedback.