The Shoppers Sparky Will Find in ChatGPT Weren’t Looking for Walmart

Walmart’s Sparky AI assistant will be embedded inside ChatGPT as soon as next week. A Walmart spokesperson confirmed that to CNBC today, the same day OpenAI formally announced it is ending Instant Checkout — the feature that let shoppers complete purchases directly inside the chatbot. The transition has been visible to anyone watching closely. At Morgan Stanley’s Technology, Media and Telecom conference on March 4, Walmart EVP of AI Acceleration Daniel Danker called Instant Checkout “a very temporary moment in time” and said: “By this time next month, you will not see that experience anymore. What you will see is that the Sparky experience will travel directly into ChatGPT and Gemini and anybody else that we ever integrate with.”

That is no longer a forecast. It is a launch timeline.

Instant Checkout Underdelivered on Every Metric That Mattered

When Walmart joined the Instant Checkout rollout last fall, the ambition was real. Walmart made roughly 200,000 products available for purchase inside ChatGPT. OpenAI expected to add items from over a million Shopify merchants. By last month, according to Forrester principal analyst Emily Pfeiffer in today’s CNBC reporting, roughly 30 Shopify merchants were actually live on the service.

The failure had two distinct causes. The first was operational: Pfeiffer described merchant onboarding as arduous and the experience as prone to errors. The second was structural. OpenAI relied partly on scraping retailer websites for product data, which meant inventory status, delivery timing, and pricing could be stale by the time a shopper saw them. “Crawling and scraping is inadequate to get the full breadth of product data that you need to do a good job of commerce,” Pfeiffer told CNBC.

Walmart’s own measurement confirmed the conversion problem. The company found that purchase rates for products sold directly inside ChatGPT ran three times lower than for products that sent shoppers to Walmart’s own site to complete the transaction, first reported by Wired. Shoppers were using ChatGPT to research. They were not buying there.

Bob Hetu of Gartner told CNBC that OpenAI “underestimated how difficult the enablement of transactions was going to be.” Walmart’s response was not to try to fix in-chat transactions. It was to stop ceding the experience to OpenAI’s infrastructure entirely.

Walmart Kept the Transaction, the Cart, and Now the Conversation

Under the embedded Sparky model, users who encounter Sparky inside ChatGPT are dealing with Walmart’s own assistant — logging into their Walmart accounts, building carts that sync across Walmart’s app and website, and completing purchases through Walmart’s own checkout. OpenAI provides the platform. Walmart owns every other layer of the interaction.

An Etsy spokesperson surfaced the data dimension of that shift in CNBC’s reporting today, in a comment that deserves more attention than it has received. The new app-based model, the spokesperson said, gives retailers “more shopper data earlier in the purchase journey” than Instant Checkout did — which only provided visibility into a transaction after it was already complete. Walmart is not simply moving checkout infrastructure. It is moving the point at which it can observe what a shopper is thinking — from the moment of purchase back into the moment of consideration.

Adobe-owned Semrush, in a survey of more than 1,000 U.S. consumers released this month, found that half of respondents had made a purchase after using AI during research, while only 22% had completed a purchase inside an AI tool. That gap — AI as discovery layer, retailer site as conversion layer — is the behavioral reality Walmart is now building its architecture around.

The Shopper Who Started With a GLP-1 Question

Danker’s March 4 comments at Morgan Stanley, reported by Digital Commerce 360 on March 19, offered the clearest picture yet of what Sparky has already found inside ChatGPT. He described watching what the earliest Sparky-in-ChatGPT users were buying, expecting they might simply replicate purchases they already make inside the Walmart app. That is not what happened.

“Top two items that people are buying from Walmart on ChatGPT: vitamin supplements and protein supplements,” Danker said. The prompts driving those purchases were not starting from commercial intent. “It’s a prompt that probably began with, I just started taking GLP-1s. What do I need to know?” A shopper seeking health guidance ends up buying a Walmart product — a purchase that would never have started with a search on Walmart.com.

That is the discovery behavior the embedded Sparky model is built to capture at scale. The items that surface in those moments are drawn from Walmart’s own catalog, not from scraped web data. Pfeiffer’s observation about scraping being inadequate for commerce is not just a diagnosis of what went wrong with Instant Checkout. It is a description of why your item’s content in Walmart’s systems now determines its visibility on platforms Walmart does not own.

Where 1P and 3P Suppliers Stand

For 1P suppliers, the transition is immediate. Sparky launches inside ChatGPT next week. The completeness and accuracy of your item content — attributes, categorization, use-case descriptions — determines whether Sparky can connect your product to a conversation that never started as a shopping trip. That has been true inside the Walmart app since Sparky launched. It now extends to a platform where, by Danker’s account at Morgan Stanley, the purchase triggers are categorically different from anything driving traffic to Walmart.com directly.

For 3P Marketplace sellers, Walmart has not specified the scope of Marketplace catalog inclusion in the initial rollout. Until there is official Marketplace communication on that point, it is reasonable to assume 1P items will have stronger early representation, since Walmart manages that catalog data directly. Pfeiffer told CNBC today that the current ChatGPT app experience “leaves a lot to be desired” and offered no guarantee the model will prove out. That caution is warranted — consumer adoption of in-chat shopping has been limited across the industry, and neither Walmart nor OpenAI has solved that cleanly yet.

What is settled is the direction. Sparky will operate inside ChatGPT, Gemini, and wherever else Walmart integrates next. The shopper who opens ChatGPT with a health question and closes it having bought a supplement never searched for a Walmart product. Whether yours was the one Sparky surfaced comes down to what your catalog says about it.

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.