Why Does the Walmart App in ChatGPT Recommend Some Products but Not Others?
The Walmart app in ChatGPT runs Walmart's own shopping assistant, Sparky, so it recommends from Walmart's catalogue using Walmart's signals, not ChatGPT's general ranking. Which products appear depends on what Walmart can deliver quickly to the shopper, the quality of each listing, its price and offer, its ratings and reviews, and, when a Walmart account is linked, the shopper's own history and cart. Separately, ChatGPT's regular shopping results, outside the app, often route buy links to Walmart but reshuffle which products they show on every run. So a product can be missing for two different reasons, and each needs a different fix.
What is the Walmart app in ChatGPT?
| Date | What happened |
|---|---|
| October 2025 | Walmart announced a partnership with OpenAI to let shoppers buy through ChatGPT with Instant Checkout |
| November 2025 | Reports say about 200,000 Walmart products became available through Instant Checkout |
| March 2026 | Walmart ended the Instant Checkout pilot and brought Sparky into ChatGPT as an app, for paid ChatGPT users first, with free users to follow |
| 2026 | A similar Walmart experience in Google's Gemini, built on Google's Universal Commerce Protocol |
Reports at the time said in-chat checkout converted at roughly a third of the rate of click-throughs to Walmart's own site, which is why Walmart moved to its own assistant, account and cart. In the app, shoppers sign in with their Walmart account, and orders go through Walmart's systems.
Why does the app show some products and not others?
Walmart has not published how Sparky ranks products, so these are the signals it is known or reported to use:
- What can reach the shopper quickly. One analysis of Sparky notes that it draws on inventory across Walmart's network of more than 4,500 US stores and fulfilment centres, so items not in stock somewhere that can reach the shopper fast may not appear. It answers “what is best that Walmart can get to you soon”, not simply “what is best”.
- Listing quality. Walmart scores listings on content, discoverability, offer, ratings and reviews. Thin titles, missing attributes and poor images make a product harder to match to a request.
- Price and offer. Where several sellers offer the same item, the offer Walmart shows depends on price, delivery promise and seller performance.
- Ratings and reviews. Sparky summarises reviews, so products with few or poor reviews have less to recommend them.
- The shopper's account. In Walmart's announcement of its Gemini experience, it said linked accounts get recommendations based on past online and in-store purchases and can combine orders with existing carts. A similar account link applies in ChatGPT, so two shoppers may see different products.
- The request itself. Sparky interprets goals, such as “stock a first apartment kitchen under $150”, and chooses products to fit the whole list.
Why do Walmart products come and go in ChatGPT's regular shopping results?
Outside the app, ChatGPT builds its own shopping results from merchant data and crawled pages, then attaches buy links from retailers. Profound's analysis of 22.5 million ChatGPT shopping results, published in March 2026, found:
- Walmart held the top buy-link position most often, at 8.78% of rank-one retailer offers, and Target had the widest overall presence.
- Results reshuffle constantly. About 95% of product titles appeared in fewer than 30% of runs of the same prompt.
- A median of five products appeared per response, ranging from one to 67.
So a product missing from one answer may appear in the next. Retailers like Walmart appear because ChatGPT routes purchases through them, not because it prefers their products. The factors behind ChatGPT's product ordering are discussed in how ChatGPT ranks shopping results.
What does Walmart's switch tell other retailers?
Walmart's move is a useful signal for any merchant planning for AI shopping. It had one of the largest catalogues in ChatGPT's in-chat checkout, yet reportedly chose to route shoppers back to its own account, cart and checkout, and OpenAI itself narrowed in-chat checkout soon after. For most retailers, that points to a practical order of work:
- Be discoverable first, through accurate product data and crawlable pages, since that decides whether you are shown at all.
- Make the hand-off smooth, with fast mobile product pages and guest checkout, since most AI shopping purchases now finish on the merchant's side.
- Treat branded apps inside assistants as an option for large catalogues, not a requirement; they need an account system, cart syncing and ongoing maintenance.
What can brands and Marketplace sellers do?
For the Walmart app:
- Raise Listing Quality: complete titles, attributes, images and descriptions in Walmart's Seller Center.
- Keep products in stock where Walmart can deliver fast, for example through fulfilment options that let Walmart promise quick delivery.
- Stay competitive on the offer: price, shipping speed and seller performance.
- Grow reviews through Walmart's own channels. Walmart asks buyers for reviews itself; sellers can add reviews through its Review Accelerator programme, approved sampling vendors, or by syndicating reviews collected on their own site through an approved provider. Do not message buyers or add package inserts asking for reviews, which Walmart's Seller Code of Conduct prohibits. Answer problems quickly.
For ChatGPT's own shopping results:
- Keep product data consistent across your site, retailers and any feeds.
- Make your own product pages crawlable, with prices and availability in the HTML.
- Sell through more than one retailer if you rely on AI discovery, since buy links rotate among them.
When AI routes shoppers to marketplaces instead of your own store, the causes are covered in why AI recommends marketplaces instead of your store.
How do you check what shoppers see?
- Test in the Walmart app in ChatGPT with realistic requests, both signed in and, where possible, from different locations, since availability shapes results.
- Test ChatGPT's regular shopping results several times per prompt and count how often your product appears.
- Compare with the Walmart mobile app, where Sparky also runs, to separate Walmart's logic from ChatGPT's. If a product is missing in both, the cause is on the Walmart side: listing, stock or offer.
We have not tested the Walmart app in ChatGPT systematically. This page is based on Walmart's announcements, news reports on the Instant Checkout pilot, a published analysis of Sparky, Walmart's seller documentation and Profound's shopping research.
What might this look like in practice?
Consider an illustrative small kitchenware brand, Gormfen, selling on Walmart Marketplace. Its pans rarely appeared when shoppers asked the Walmart app in ChatGPT for a starter cookware set, though they ranked well in Walmart search for exact names. Its listings lacked material and hob-compatibility attributes, and it shipped from one warehouse with five-day delivery. Gormfen completed the attributes, rewrote titles to state the set contents, and moved its best sellers into faster fulfilment. Over the following weeks the sets began appearing in Sparky's suggestions for kitchen starter requests, mostly for shoppers in regions with next-day delivery.
What do people ask about the Walmart app in ChatGPT?
Is the Walmart app the same as ChatGPT Shopping?
No. The app runs Walmart's own assistant and catalogue; ChatGPT Shopping shows products from many merchants.
Can I pay to appear in the Walmart app's recommendations?
Walmart has not said how, or whether, sponsored products appear in Sparky's answers in ChatGPT.
Why do results differ between shoppers?
Availability near the shopper and linked account history change what Sparky recommends.
Does Walmart still use Instant Checkout in ChatGPT?
No. It ended the pilot in March 2026 and moved to its own app and checkout.
Published by The Kovmere Team · updated Sep 29, 2026 · More resources