Walmart on October 30, 2026, added six capabilities to Sparky, the AI shopping agent it runs in its app, on its website and in its stores, in a newsroom post timed for the holiday season. Shoppers in the United States can now photograph a handwritten list, or export one from the Notes app, and have Sparky turn it into a shopping cart. They can also upload a picture of a stained sweater and get product recommendations, make a shelf label flash in a store, have Sparky write recipes and load the ingredients into the cart, ask for healthier substitutes and converse in Spanish or French. The post, titled "Sparky Makes Holiday Magic Real" and filed under Innovation, carries no executive quotes, no usage figures and no mention of advertising. A footnote says voice capabilities are "expected to launch chainwide by the end of the year."

In Short

Walmart's shopping chatbot, Sparky, can now look at a photo of a stain or a handwritten shopping list and turn it into things to buy, and in stores it can make a price tag on the shelf light up so you can find an item. That matters to brands because Walmart's own AI increasingly decides which product lands in the basket when a list only says "paper plates" or "marshmallows." Walmart has not said whether any of Sparky's picks are paid ads, and talking to Sparky out loud is due across the chain by the end of the year.

Six additions, framed around the holiday list

"The holidays have a way of turning one shopping list into five," the post begins. It goes on to describe Sparky as "Walmart's AI shopping agent" and says it works "with the same Every Day Low Prices customers already see in-store, online and through the Walmart app." Customers "can type, talk* or share an image with Sparky," according to Walmart, and the asterisk on "talk" leads to the footnote about voice.

The post presents "six new ways Sparky can help." Two of them change how a shopper can address the assistant: by image, and in a language other than English. The other four turn a request into a basket, or send the shopper to a specific spot on a shelf. Each one ends in the same place, a product module chosen by Sparky.

A photo of the problem

The first feature takes a picture as the query. Walmart's examples are hot cocoa spilled on an "ugly sweater," cranberry sauce on a tablecloth and an "inspo pic" of a look a shopper wants to copy. "Snap a photo or take a screen grab, upload it, and Sparky can identify what you need, recommend the right products, and help get them delivered in as little as 30 minutes," according to Walmart.

The app example Walmart published with the feature shows the order of events. Under a photo of a stained knit, the user writes: "Help me clean this hot chocolate stain on my favorite sweater!" Sparky answers with a method before any product card appears - a color-safe pre-treatment spray plus a gentle detergent - and warns the user to test the spray on a hidden spot first, "since the red and green yarn could bleed." Only then does a module headed "Stain Remover - Sparky's Picks" appear, its first card labelled "Color-safe formula, gentle on patterned knits."

Visual queries are already a large shopping channel elsewhere. Google has put Lens at close to 20 billion visual searches a month, roughly 20% of them shopping-related. On Google, a photographed stain can lead to any retailer. Inside Sparky, the answer is a Walmart product.

Handwritten lists into carts

The second feature is the one Walmart describes with the most enthusiasm. "Just take a picture of your handwritten list or export your digital list from the Notes app and watch Sparky do all the work," according to the post, which names gift lists, party supply lists, "lists about lists" and "your children's handwritten Christmas lists" as candidates.

Sparky's reply in Walmart's own example says more than the description does. "Here's your holiday to-do list, matched item by item," it reads. "The Glad trash bags you've bought before are out of stock right now. This shopping list is for your usual store: Rogers Pleasant Crossing Blvd Supercenter. I recommended options for items you haven't purchased before." The resulting list holds nine items totalling $64.21, led by a $16.99 set of 100 Christmas paper plates and napkins.

Three mechanisms sit inside those four sentences. Sparky reads the shopper's purchase history, so a product bought before becomes the default match for a word on the list. It checks stock at one specific store, which it calls the shopper's "usual store." And where no purchase record exists, it chooses a product itself. Purchase history is Walmart's own first-party data, and here it decides which brand fills an unbranded line. Who picks the paper plates? For a brand, the difference between "items you've bought before" and "items you haven't purchased before" is the difference between being matched automatically and competing to be recommended. The post does not explain how Sparky ranks the options it recommends for items with no history.

Walmart is not first with this function. Amazon added handwritten list recognition to Rufus on iOS in November 2025, as one of more than 50 technical upgrades, with an Android version scheduled for later.

Shop to Light

The third feature moves Sparky from the screen to the shelf edge. Under the name Shop to Light, a shopper searches for an item in Sparky, uses a pin on the store map to find the aisle and, once there, taps "Flash price tag." The shopper can then "watch the shelf label light up, helping you spot exactly what you're looking for," according to Walmart. The image that accompanies the feature shows a shelf label glowing blue beneath a row of jars.

The post gives no store count, no rollout date and no description of the shelf hardware. Nor does it say whether the function covers every product in a store or only items on labels that can light. What it does establish is a link between the assistant and a physical position on a shelf - the same territory that in-store retail media sells through screens, audio and connected shopping tools. A search that starts in the app now ends at a lit label in the aisle, with Walmart's software controlling both ends.

"Hosting your first Friendsgiving for a group of 15, on a budget, and 2 guests are vegetarian? Tell Sparky your culinary vision and it generates step-by-step recipes and ready to buy ingredients," Walmart writes. The post adds, with some candour, that "while it unfortunately can't help with the cooking, Sparky can add all the ingredients you need right to your cart when you are ready to go."

The app example puts a figure on the budget. The prompt reads: "I'm hosting Friendsgiving for 15, with two vegetarians. Budget is $125. Help me plan the menu!" Sparky replies that it "found no close matches in your purchase history, so I built this from scratch," then sets out its reasoning: "Turkey is the budget anchor (about $1.78/serving), so scale sides to fit your $125." Two recipe cards from a source labelled Recipe Hub follow. One is an Herb Roasted Whole Turkey at an estimated $1.78 per serving, with 20 minutes of preparation and two hours of cooking; the other is a Vegetarian Chestnut Root Vegetable Casserole at an estimated $4.65 per serving, with 35 minutes of preparation and one hour and 10 minutes of cooking. Sparky also notes that the turkey recipe uses chicken broth and suggests vegetable broth for anything the vegetarians will share. "Open a card to add its ingredients," the reply ends.

Purchase history shows up here too, as the first place Sparky looks before it generates anything new. The recipe cards themselves are a commercial step: each one opens into a list of ingredients that Sparky can add to the cart at once.

Healthier swaps

The fifth feature works on two levels. "Sparky can swap in healthier ingredients for your favorite recipe, or suggest a better-for-you version of something already sitting in your shopping cart," according to Walmart. The sample requests are "What's a healthier swap for my mom's turkey stuffing?" and "Find me a lower-sugar alternative to these marshmallows."

In the app example, asked for "three healthier swaps to classic Thanksgiving dishes that still taste great," Sparky proposes homemade cranberry sauce in place of canned, cauliflower mash alongside or instead of mashed potatoes, and a lower-sugar stuffing base. It cites a regulator for its definition: "Per the FDA, 5% DV or less of added sugar per serving is low." The first product card is a 12-ounce bag of fresh cranberries at $0.76, marked down from $0.97, with 2,972 ratings and the note "Simmer with orange for sauce; 4 g fiber, no sodium."

The second level is the one brand owners will read closely. Suggesting a different version of an item "already sitting in your shopping cart" means Sparky can propose replacing a product the shopper has already chosen. The post does not say how replacement candidates are selected, whether the replacement can come from a different brand, or whether any placement in that slot is paid.

Spanish and French

The sixth feature carries a qualifier in its heading: "also available in Mexico and Canada." "Simply type or speak in either language to Sparky, it detects the natural language input and responds in-kind, ensuring effortless assistance," according to Walmart. The post says the feature "breaks down language barriers for millions of shoppers," without a figure.

Walmart's example is in Spanish. The prompt reads: "Quiero decorar mi casa para Navidad sin gastar mucho. Ayúdame a encontrar luces, adornos y detalles festivos por menos de $50." (I want to decorate my house for Christmas without spending much. Help me find lights, ornaments and festive details for under $50.) Sparky answers in Spanish that it found items "casi todos por menos de $15" (almost all under $15). The product module is headed "Selección de Sparky." Its first card, a 100-light warm LED string at $5.88 with 2,590 ratings, carries the badge "En 500+ carritos de personas" (in 500+ people's carts). The surrounding interface, including the "Ask me anything" field, stays in English. The option to speak in either language depends on the voice rollout described in the footnote.

Features already in place

Walmart presents the six additions as building on gift functions shoppers already use. Sparky "can suggest gifts based on their personal interests, hobbies, age range and budget," according to the post, and "can even show what's in stock at a local store and available for delivery in as fast as 30 minutes." That delivery figure appears twice in the post, and it is not new: PPC Land reported in February that Sparky could promise delivery in under 30 minutes in many markets.

What the post leaves out

The availability sentence is broad. "Available across Walmart customers in the U.S., and scaling to Canada and Mexico and Sam's Club members in the U.S., Sparky is spreading holiday joy to millions of shoppers," according to Walmart. "Scaling" comes without a date for any of the three groups, and the post's subheading says only that "Access expanded to millions more" shoppers.

Voice is the clearest gap. The post's opening section lists talking as one of three ways to use Sparky, yet the footnote says voice is only "expected to launch chainwide by the end of the year." Walmart does not define "chainwide," which could refer to stores, the app or both.

The subheading also refers to "multi-model features available in the U.S." The post does not name the large language models involved or say whether "multi-model" refers to several AI models or to the several kinds of input - text, images and voice - the features accept. PPC Land reported in March that Sparky runs on open-source generative AI models alongside retail-specific models.

Data handling goes unmentioned. Several of the features depend on images of a shopper's home, clothing and handwriting, including, by Walmart's own example, children's Christmas lists. The post does not say how uploaded images are stored, how long they are kept or whether they inform later recommendations. It also gives no user numbers, no per-feature launch dates and no store count for Shop to Light.

Why it matters for advertisers

Walmart is a media owner as well as a retailer, and Walmart Connect is its retail media network. Sparky has been treated as potential ad inventory for at least a year. Walmart tested advertising formats within Sparky during fall 2025, which PPC Land described in January as the company's first integration of advertisements into AI shopping experiences. The same report cited an August 2025 survey in which 81% of surveyed Walmart customers said they used Sparky to check product availability and review specifications.

Use has grown since. In February, Walmart said about half of its app users had engaged with Sparky and that their average order value was approximately 35% higher than that of other customers. By August, the number of customers using Sparky was up 70% year over year, and Walmart put the gap at 40% more spent per order. Walmart Connect's US advertising grew 43% in that quarter, excluding VIZIO.

The October 30 features move more of the brand decision inside Sparky. A handwritten line such as "paper plates" names a category, not a brand; a recipe calls for broth; a swap replaces something already in the cart. In each case, Walmart's examples end in a module labelled "Sparky's Picks" or "Selección de Sparky," and none of the product cards visible in those examples is marked as sponsored. Whether paid placements appear in these modules, how they would be labelled, and how they would rank against the purchase-history matches the list feature relies on are questions the post does not address.

Competitors have moved further on pricing this kind of placement. Amazon's AI shopping prompts became billable at cost per click on March 25, 2026, after a free beta that began in November 2025, and on May 13 Amazon merged Rufus and Alexa+ into Alexa for Shopping, citing more than 300 million Rufus customers in 2025 and nearly $12 billion in incremental annualized sales. Google introduced shopping ads inside AI Mode on February 11, 2026, labelled as sponsored, and on October 15 began rolling out Universal Cart across US Search and the Gemini app.

Walmart has spent the year pushing Sparky outward as well as building it up. In January it announced checkout inside Google's Gemini app and AI Mode alongside Target. In March, after a Walmart executive said products sold inside ChatGPT converted three times lower than click-outs to Walmart's site, the company moved to run Sparky inside ChatGPT, with a Gemini version to follow. On October 21, Walmart said its Agent Gateway would connect outside AI agents through Sparky and that agents would see the same prices as human customers. The October 30 post does not say whether image reading, list matching or swaps apply when an outside agent reaches Walmart through that gateway.

The holiday framing is not incidental. Two days earlier, Walmart's own trend report found that 43.8% of its panel members named affording the holidays as their top stressor; that report did not mention Sparky by name. The Sparky post returns to money again and again: a Friendsgiving menu scaled to $125, decorations for under $50, gifts filtered by budget, cranberries at 76 cents. Each of those examples is also a moment at which Sparky, not the shopper, names the product.

Timeline

Summary

Who: Walmart, for customers using Sparky in the US, with expansion described as "scaling" to Canada, Mexico and Sam's Club members in the US. Brands selling at Walmart and advertisers buying through Walmart Connect are affected by where Sparky's product choices land.

What: Six additions to Sparky: image input for problems such as stains, conversion of handwritten or exported lists into carts, Shop to Light (a store map pin plus a "Flash price tag" button that lights the shelf label), recipe and ingredient generation for a described meal, healthier swaps for recipes and cart items, and conversation in Spanish and French. Walmart's examples show Sparky using purchase history and a shopper's usual store, and presenting products in modules labelled "Sparky's Picks." No advertising, usage figures, model names or data-handling terms are given.

When: Announced on October 30, 2026. Voice capabilities are expected to launch chainwide by the end of 2026. No dates are given for the expansion to Canada, Mexico or Sam's Club.

Where: The Walmart app, Walmart's website and Walmart stores in the US; Spanish and French are also described as available in Mexico and Canada.

Why: Walmart positions the features as help with holiday shopping that takes "a little more thought, a few more steps and time that's already in short supply." For the marketing community, the update moves more brand selection into Walmart's own assistant, which Walmart Connect has already tested as ad inventory, at a time when Amazon and Google charge for placements inside their AI shopping surfaces.