ChatGPT's shopping feature is showing a clear preference: products from merchants who submit structured data feeds are surfacing more often than products pulled from general web pages. If your store doesn't feed OpenAI clean, structured product data, your visibility in ChatGPT's shopping results may already be shrinking.
When OpenAI rolled out shopping features inside ChatGPT earlier this year, the system pulled from a mix of sources — crawled web content, retailer partnerships, and structured feeds submitted directly by merchants. Recent tracking shows the balance tipping hard toward that last category. Products with feed data — price, availability, images, identifiers like GTINs — are showing up more consistently than products ChatGPT has to infer from a webpage's text and markup.
This matters because a product feed isn't something every small business has. It's a structured file, often in XML or CSV format, that lists every item a store sells along with standardized attributes. Big platforms like Shopify, WooCommerce, and BigCommerce can generate these automatically or through plugins, but a merchant running a custom-built site or a barebones storefront may not have one at all.
The practical effect is a widening gap between stores that have invested in feed infrastructure and those that haven't. A shop with a well-maintained feed can appear in ChatGPT's product picks even if its website content is thin. A shop with excellent product pages but no feed may become nearly invisible to the same tool, regardless of how good its content is.
This is not a new playbook. Google Shopping went through the same evolution over a decade ago, moving from a free listing model based partly on organic crawling to one almost entirely dependent on Merchant Center feeds. Bing followed a similar path. What's different this time is the speed: ChatGPT Shopping is only months old, and it's already showing the feed-dependency pattern that took Google years to fully establish.
The broader trend across AI shopping tools — including Perplexity's shopping features and Google's AI Overviews with product panels — is a move away from open-web discovery toward structured, merchant-supplied data. AI systems appear to trust standardized feeds more than they trust parsing a website's HTML, likely because feeds reduce ambiguity about price, stock, and specifications.
For small businesses, this creates a fairly concrete task list rather than a vague warning. If you sell products online, check whether your e-commerce platform already generates a product feed — most major platforms do, often buried in settings under "sales channels" or "marketing." If you're on a custom site, you may need a feed management tool or a developer to build one.
The cost of ignoring this isn't abstract. Feed-based visibility increasingly determines whether your products show up when someone asks ChatGPT for shopping recommendations, a channel that's small today but growing as more consumers use conversational search before searching Google directly. Businesses that already run Google Shopping campaigns have a head start, since the same feed can often be adapted for both.
There's also a data-quality dimension. A feed with missing images, wrong pricing, or unclear availability can hurt more than having no feed, since it may surface incorrect information to a customer using ChatGPT to make a buying decision. Accuracy matters as much as existence.
Watch whether OpenAI publishes formal feed specifications or a merchant portal similar to Google Merchant Center, which would confirm this is a permanent structural feature rather than a temporary weighting quirk. Also watch whether OpenAI begins charging for feed placement or prioritized listings, following the path Google Shopping took from free to pay-to-play.
The bottom line: if you sell products online, check this week whether your store has a working product feed and whether the data in it is current. Visibility in ChatGPT Shopping is starting to depend less on your website and more on that file.