Walmart’s Sparky Agent Is Lifting Orders by 35%: The Agentic Commerce Wake-Up Call Brands Can’t Skip

Here is a number worth sitting with: Walmart’s Sparky AI shopping agent is generating average order values 35% higher than non-Sparky customers. That figure landed during Walmart’s Q1 FY2027 earnings call on May 21 — and it hits differently depending on where you’re sitting.

If you’re a brand manager, it means the customers most likely to spend more are now primarily shopping through an AI interface, not a search bar. If you’re a COO weighing the agentic commerce question, it means the infrastructure is live and operating at scale at one of the world’s largest retailers. Not a pilot.

A 35% Gap That Compounds

Walmart US CEO John Furner described a retailer actively “becoming AI native.” The Sparky metrics back that up with unusual precision. Weekly active users grew over 100% year-on-year in the quarter. Units purchased through the agent rose more than fourfold versus the prior period. And the intelligence of the agent itself improved by 40% year-to-date, according to the company.

What’s striking about the AOV gap isn’t just its size — it’s the dynamic it implies. A shopper who uses an AI agent, gets personalised recommendations, and completes a session doesn’t just spend more once. They return with a different expectation of how retail works. The platform trains them; they train the platform. Brands that surface well inside Sparky’s recommendation logic are gaining premium, repeat shoppers. Brands that don’t appear there are invisible to a growing and increasingly loyal customer segment.

Walmart’s enterprise e-commerce grew 26% in Q1, with U.S. marketplace sales up almost 50%. For the brands in that marketplace, the implications are direct. According to DigitalCommerce360, Sparky now spans the app, web, and in-store experience — covering personalised replenishment, meal planning, and inventory-aware recommendations that factor in delivery speed.

Three weeks ago, Amazon retired Rufus, folding its AI shopping capabilities into a revamped Alexa for Shopping. That transition raised urgent questions about brand visibility in AI-mediated search. Walmart’s Sparky, meanwhile, posted growth numbers that suggest it’s accelerating, not consolidating.

Epinium data

Across the 400+ brands managed through Epinium Platform, those that complete AI-assisted product content enrichment on their top 50 Walmart SKUs within the first 30 days of onboarding see catalogue quality scores rise by an average of 38 points — the same completeness signals that feed directly into shopping agents’ recommendation ranking.

The Super Agent Your Brand Team Hasn’t Heard Of

Consumer-facing Sparky gets the headlines. But the more structurally significant development may be Marty — Walmart’s partner super agent, designed specifically for suppliers, advertisers, and marketplace sellers. Walmart is building four super agents in total: Sparky for consumers, Marty for commercial partners, a third for store associates, and a fourth for internal tech teams.

Marty’s existence matters because it signals that Walmart’s AI architecture isn’t just improving the shopping experience for end customers — it’s remaking the operating relationship between the retailer and the brands it stocks. When an AI agent manages inventory signals, promotional calendars, and seller communications on Walmart’s side, the brands that thrive will be those whose data and content are clean enough to interface with that agent clearly.

This is the shift most brand teams haven’t fully registered. You can’t negotiate with Marty the same way you’d call a category buyer. What’s needed now is structured data, consistent content, and fast response to algorithmic signals — an operating model that rewards brands already running on AI-assisted infrastructure.

Is your brand ready for the agent layer? Discover how Transform builds the AI-ready commercial infrastructure your retail partners now expect →

What we’re seeing at Epinium is that the gap between AI-ready brands and everyone else is widening faster than it was twelve months ago. Google unveiled its Universal Cart at I/O last week, raising the same structural question: when an AI agent intermediates the purchase, who is your brand actually communicating with?

What “AI Native” Actually Demands

For brands selling through Walmart, “AI native” has a precise operational meaning. Products need to be machine-readable at every layer: titles structured for retrieval, attributes complete, imagery processed for visual AI, pricing signalled programmatically. Sparky’s recommendations are shaped by content quality, relevance signals, review velocity, and fulfilment performance — all factors that AI-driven catalogue management can systematically improve.

Sub-three-hour delivery orders at Walmart grew over 60% year-on-year in Q1. Sparky isn’t just a discovery layer — it’s increasingly tied to fulfilment speed signals. A product with rich, accurate content and fast fulfilment has a structural advantage inside an agentic system that a human browser would never notice, but that Sparky processes on every query.

The brands gaining ground inside platforms like Sparky are not the ones optimising after the fact. They are the ones whose content and commercial processes were already built for this moment.

What exactly is Walmart Sparky?

Sparky is Walmart’s AI shopping agent, live across the app, web, and in-store surfaces. It delivers personalised product recommendations, replenishment suggestions, and meal planning assistance. Customers using Sparky spend approximately 35% more per order than non-users, and weekly active users grew over 100% in Q1 FY2027. Units purchased via Sparky rose fourfold in the same quarter.

Does the 35% AOV premium prove Sparky causes more spending — or do high-spenders simply use it more?

Almost certainly both, and that distinction matters. There is a selection effect: engaged shoppers are more likely to try an AI agent. But a fourfold increase in units through Sparky in a single quarter, combined with 100%+ user growth, signals the platform is reaching mainstream shoppers, not just power users. As penetration broadens toward the general app population, any selection bias shrinks.

What is Walmart’s Marty agent, and why should brand teams care?

Marty is Walmart’s super agent for commercial partners — suppliers, advertisers, and marketplace sellers. Unlike Sparky, which faces shoppers, Marty sits between Walmart’s buying function and the brands that supply it. In practice, it means vendor-retailer operations will increasingly run through structured AI interfaces rather than account management calls, rewarding brands with clean, consistent, machine-readable data.

How should a brand optimise product content to perform better inside Sparky’s recommendations?

Prioritise complete and consistent product attributes, accurate inventory data, and competitive review velocity. Because Sparky integrates meal planning and lifestyle context, brands in food, personal care, and home categories should align content to use-case signals — not just SKU-level specs. Fast fulfilment is also weighted: Sparky surfaces products that can be delivered quickly alongside relevant content.

When is Sparky optimisation not the right priority?

If your brand has minimal Walmart marketplace presence, fix distribution before worrying about agent visibility. And if your catalogue has fundamental content gaps across multiple retailers, addressing those systematically will outperform any single-platform tactic. Sparky optimisation multiplies an already-solid retail content foundation — it doesn’t substitute for one.

Walmart’s Q1 numbers represent something harder to dismiss than a press release. They show that agentic commerce is producing measurable commercial outcomes at scale — and that the benefit isn’t distributed equally. Brands inside the algorithm capture premium customers. Those outside it are simply absent from a growing share of Walmart’s highest-value transactions. The architecture is set. The only remaining question is who is building for it.

Ready to build AI-ready brand infrastructure? The commercial layer between brands and retailers is becoming algorithmic — at Walmart, at Amazon, and increasingly everywhere else. Epinium Transform helps you design and deploy the AI-assisted processes that ensure your brand is visible where it matters most. Book your free 30-minute AI diagnosis →