By
Austin Coker
Reading time: 15 min
Last updated: November 2, 2026.
Walmart Sparky is the generative AI shopping assistant inside the Walmart app and website. Shoppers type, photograph or (soon) speak a request, and Sparky answers with a short list of products pulled from Walmart’s catalog, ready to add to the cart. Walmart describes it as a tool that helps people “find and compare products, build lists, get personalized recommendations and plan for occasions by synthesizing reviews and product data” (source: Walmart Technology).
Walmart hasn’t published a ranking formula for Sparky, and it may never do so. Its announcements do show what Sparky reads: your product data, your reviews, your price, whether the item is in stock nearby, and the shopper’s own purchase history. Those are the levers a seller controls, and most of them are the same ones Walmart already scores in Seller Center.
This guide covers what Sparky is, what it can do as of November 2026, how it ends up recommending one product over another, how it connects to ChatGPT and Gemini, where ads fit in, and what to change on your listings. Sparky changes quickly, so we plan to refresh this page every month.
Sparky is Walmart’s AI shopping agent. It launched across all categories in the Walmart app in June 2025 behind an “Ask Sparky” button (source: Walmart, June 6, 2025). It now also takes photos, handwritten lists and Spanish or French requests, guides shoppers to the shelf in stores, and runs inside ChatGPT and Gemini.
Sparky recommends products by combining:
Walmart reported on August 20, 2026 that the number of customers using Sparky was up 70% from a year earlier, and that Sparky users spend 40% more per order than customers who don’t use it (source: Walmart Q2 FY27 earnings call transcript).
For sellers, the practical work is:
Sparky is a conversational assistant built by Walmart. Walmart introduced it on June 6, 2025 as “a GenAI assistant that synthesizes reviews, offers occasion-based recommendations, and helps customers plan, compare and purchase with confidence.” At launch, it could search for items, summarize reviews, answer product questions and compare options. Walmart also said Sparky would later reorder household essentials, book services and understand text, images, audio and video (source: Walmart, June 6, 2025).
It isn’t one model. Daniel Danker, who oversees design and product for Walmart, told WIRED that Sparky combines open source generative AI models with retail-specific models trained on Walmart data. Questions are routed to different models depending on which gives better answers (source: WIRED, March 18, 2026). Walmart’s own description mentions “multi-agent orchestration, fallback handling and emerging voice and camera capabilities” behind the scenes (source: Walmart Technology).
Sparky is available to Walmart customers in the U.S. and, per Walmart, is “scaling to Canada and Mexico and Sam’s Club members in the U.S.” (source: Walmart, October 30, 2026). Walmart Mexico launched its version in the Walmart app on September 24, 2026 (source: Walmart México, September 24, 2026).
People search for it as “Sparky AI Walmart” or “Walmart Sparky AI.” They’re the same product.
Walmart added six capabilities on October 30, 2026, ahead of the holidays. Here’s what Sparky does now, based only on Walmart’s announcements:
Two of those matter more to sellers than the rest. Photo input means a shopper can skip the search bar entirely: they show Sparky a stained sweater and get a stain remover. Recipe and swap requests mean Sparky picks products against constraints like “vegetarian,” “for 15 people” or “lower sugar.” If your listing doesn’t state those facts, Sparky has nothing to match against.
Walmart hasn’t published how Sparky ranks products. It also hasn’t said whether it treats items Walmart sells itself differently from Marketplace items. What its announcements and executive comments do show is the set of inputs.
Sparky answers by “synthesizing reviews and product data” (source: Walmart Technology). Product data means the record you control: product type, attributes, title, description, key features and images. Walmart’s own Marketplace guidance says categorization and attributes are “integral to determining how your items will appear in search and browse results,” and that without attributes “customers will likely have a hard time finding your items” (source: Walmart Marketplace Learn).
Sparky’s newer features are built around requests with conditions. Walmart’s examples include a Friendsgiving for 15 people on a budget with two vegetarian guests, a “lower-sugar alternative to these marshmallows,” and gifts filtered by “personal interests, hobbies, age range and budget” (source: Walmart, October 30, 2026). A product can only qualify for those answers if the matching facts are on the listing.
On the August 2026 earnings call, CEO John Furner described a shopper who asked for a high-protein weekly meal plan. He said Sparky “even recognized the ingredients they had recently purchased, both online and store, so they didn’t buy something they already had” (source: Walmart Q2 FY27 earnings call transcript). Personal history favors the brands a shopper already buys. That makes repeat purchase and first-order experience more valuable than before.
Sparky “can even show what’s in stock at a local store and available for delivery in as fast as 30 minutes” (source: Walmart, October 30, 2026). Walmart hasn’t said whether stock or speed affects ranking. But an assistant that promises delivery in 30 minutes has little reason to lead with an item that ships in a week.
Walmart Connect tested its first ad formats tied to Sparky in fall 2025 and said it would continue in 2026 (source: Walmart Connect, January 6, 2026). More on that below.
What Walmart hasn’t said is just as useful. It hasn’t said Sparky uses the Listing Quality Score, how much reviews weigh against price, or whether Pro Seller status matters. Be skeptical of anyone who claims to know the exact weighting.
Yes. Walmart now lists ChatGPT and Gemini among its “conversational commerce integrations,” alongside Sparky in its own app and Meta’s Muse (source: Walmart, October 21, 2026). The ChatGPT relationship has already changed shape once.
In the new setup, a shopper signs in to Sparky the first time they meet it in ChatGPT, and their Walmart cart syncs across the app, the website and ChatGPT (source: WIRED).
The Gemini experience works a little differently. Walmart said on January 11, 2026 that Gemini “will automatically include Walmart and Sam’s Club in-store and online products when it’s relevant.” Customers who link accounts get recommendations for complementary items based on past purchases. The experience is built on the Universal Commerce Protocol (source: Walmart, January 11, 2026). If UCP is new to you, read our explainer on what the Universal Commerce Protocol is.
The upshot for sellers is that “walmart chatgpt” and “walmart gemini” are the same catalog showing up in more places. Fix the product record once and it travels.
On October 21, 2026, Walmart also announced Agent Profiles, a Walmart Agent Gateway that uses MCP, and terms for third-party AI agents. Customers can set spending limits such as $50, $100 or $200 per agent, and purchases need approval by default (source: Walmart, October 21, 2026). Walmart says it prices “products, not people,” so agents see the same prices shoppers do.
Yes, in test form. Walmart Connect said in January 2026: “In fall 2025, we tested our first advertising formats tied to Walmart’s Sparky agent, which we’ll continue this year.” In the same post, it cited a Walmart customer survey from August 2025 in which 81% of surveyed customers said they had used Sparky to check product availability and review product specifications before buying (source: Walmart Connect, January 6, 2026).
Walmart hasn’t published a self-serve Sparky ad product, pricing or setup steps on its public pages. If Sparky placements matter to your category, ask your Walmart Connect contact which formats your account can test. Keep the organic work going either way. The same product record feeds both the organic answer and the ad.
Walmart has shared several usage figures. Each comes from a different date and a different speaker, so compare them with care:
Walmart hasn’t published a total user count. “Spend more per order” is a correlation, not proof that Sparky causes the difference. People who plan a whole party in one chat were always going to buy more than someone grabbing milk.
Sparky’s closest rival is Amazon’s Alexa for Shopping, the assistant that replaced Rufus and merged it with Alexa+ (source: Amazon). We cover it in detail in our guide to Alexa for Shopping, and the older version in Amazon Rufus.
If you sell on both marketplaces, most of the work overlaps: complete attributes, answer-first copy, honest reviews and reliable stock. The main difference is reach. Sparky also follows the shopper into a physical store and into third-party chatbots.
None of these steps guarantees a recommendation. Each one gives Sparky more accurate facts to match against a shopper’s request.
Choose the right product type first, since it decides which attributes you can fill in. Then complete every attribute a shopper might filter on: size, count, material, age range, dietary claims, compatibility. Walmart says high-quality attributes “can also lead to better product discoverability and visibility” (source: Walmart Marketplace Learn).
Sparky gets questions like “nut-free snacks for a class party under $25” more often than “snack bars.” Put the facts that answer those questions in the title, the key features and the description: who it’s for, what it’s free of, how many it serves, what it works with. Use the same wording in your attributes so nothing conflicts.
Review synthesis has been part of Sparky since launch. If a recurring complaint shows up in your reviews, expect it to show up in Sparky’s summary. Fix the product or packaging issue behind it, answer it in your listing copy, and follow up with buyers through Walmart’s approved channels.
Sparky shows local store stock and delivery in as little as 30 minutes. Out-of-stock and slow-shipping items are weak candidates for that kind of answer. Walmart’s Listing Quality guidance already lists expedited shipping and competitive prices among the recommendations it gives sellers (source: Walmart Marketplace).
Walmart hasn’t said Sparky uses the Listing Quality Score. Still, the score “considers content, discoverability, offer, ratings, and reviews,” which are the same facts Sparky reads (source: Walmart Marketplace). It’s the closest official health check you have. Start with your top 20 items by revenue.
Sparky now accepts photos from shoppers and recommends products from them. High-resolution images that show the product, its size and its use give the system more to work with. They also help the shopper confirm Sparky’s pick.
Ask Sparky the 10 to 20 questions your buyers really ask, including budget, occasion and “healthier” versions. Log which products appear, in what order, and what Sparky says about yours. Repeat monthly, and ask in Spanish too if Hispanic shoppers are a big share of your buyers. This is the only reliable way to see your Sparky visibility, since Walmart doesn’t report it.
Sparky now runs inside ChatGPT and Gemini, and Gemini can surface Walmart products on its own. Shoppers often research a product elsewhere before they reach Walmart. Make sure your brand site, your Walmart listing and third-party reviews describe the product the same way. That’s the core of generative engine optimization.
We run prompt audits across AI shopping assistants and show you where your products show up, where competitors win, and what to fix first.
Sparky changes fast. Here are the dated milestones from Walmart’s own announcements and reporting on them:
Sparky is one of several assistants that now pick products for shoppers. The others include Alexa for Shopping, ChatGPT, Gemini and Meta’s Muse. We compare them in AI shopping agents compared and cover the bigger picture in our guide to AI shopping assistants. They all share one habit: they recommend from structured product data, reviews and what the wider web says about you, then show a short list instead of a page of results. Your product often has to come up in more than one of those systems before a shopper buys it.
That’s why the work above isn’t only “Walmart SEO.” Clean attributes help Sparky and Google Shopping. Answer-first copy helps Sparky and ChatGPT. Third-party mentions help Gemini and Sparky-in-Gemini.
At 95 Projects, we help ecommerce brands get recommended across Google, ChatGPT, Gemini and retailer assistants like Sparky. We pair listing and site content with the third-party coverage AI systems rely on. For Buoy, that meant going from not showing up at all to ChatGPT’s #1 recommendation for “best electrolytes with no sweeteners” in three weeks (case study). If you want to see where your products stand in AI shopping answers today, look at our GEO for ecommerce service or book a call.
Austin Coker is the CEO and founder of 95 Projects. He has worked in SEO since 2020, when he built and sold an ecommerce brand that grew mainly through search. He founded 95 Projects in 2022 to help SaaS, ecommerce, and B2B service companies show up wherever their buyers search, from Google to ChatGPT and AI Overviews.
Have questions about working with us? Book a 30-minute strategy call to discuss your goals and see if we’re a good fit.
No-pressure conversation
Clear next steps (if we’re a fit)
Talk directly to an expert
Walmart Sparky is Walmart’s generative AI shopping assistant in the Walmart app and website. It finds and compares products, summarizes reviews, builds carts from lists and photos, plans meals and occasions, and personalizes suggestions based on your purchase history. Walmart launched it in June 2025.
Walmart hasn’t published a ranking method. Its announcements show that Sparky uses Walmart’s product data and reviews, the constraints in the shopper’s request, the shopper’s past purchases, and local stock and delivery speed. Sponsored ad formats are also being tested.
They’re now connected. Walmart first sold about 200,000 items through ChatGPT’s Instant Checkout, then moved to a Sparky-powered Walmart app inside ChatGPT in March 2026. There, shoppers link their Walmart account and check out with Walmart. Gemini has a separate Walmart integration built on the Universal Commerce Protocol.
Walmart Connect has tested ad formats in Sparky since fall 2025 and said it would continue in 2026. Walmart hasn’t published a self-serve Sparky ad product or pricing on its public pages, so ask your Walmart Connect contact what your account can access.
Walmart hasn’t said that Sparky excludes Marketplace items or treats them differently. Sparky recommends from the Walmart catalog, so a complete, in-stock Marketplace listing is a candidate like any other. Test your own category prompts to confirm.
Yes, in some markets. Walmart Mexico launched Sparky in its app on September 24, 2026. Walmart said on October 30, 2026 that Sparky is scaling to Canada and Mexico and to Sam’s Club members in the U.S., with Spanish and French support.
Often. Walmart has announced new Sparky capabilities several times in 2026 alone, most recently on October 30. Check Walmart’s newsroom before you change listings based on any guide, including this one. We update this page monthly.
