ChatGPT is no longer just a chatbot. For several months, the Shopping feature has been recommending products directly in conversation, complete with images, prices, and purchase links. Not like a traditional search engine returning dozens of results, but like a consultant answering „which phone should I get under $500?” with three or four specific options, comparisons, and reasoning.
For users, the experience is straightforward: ask a question, get tailored recommendations. But for brands selling online, what happens behind those recommendations matters enormously. Search Engine Journal recently published data from Profound, an analytics firm specializing in AI shopping, showing one clear finding: 65% of product recommendations now come from integrated feeds, not web search. If your products aren't in a feed connected to the platform, their chances of appearing in these conversations are slim and shrinking.
This isn't an algorithmic penalty. It's an infrastructure decision. And the implications for anyone selling online are immediate.
On July 10, the rules changed overnight
Profound monitors ChatGPT Shopping recommendations daily. Across a sample of 1.76 million prompts analyzed in July (50% of total daily volume), the data shows a sudden shift. On July 10, with the GPT-5.6 update, feed-integrated product sources jumped from 8.26% to 61.54% of all recommendations. In a single day.
What does that mean in practice? Before July 10, ChatGPT crawled the web to find products. It read product pages, compared descriptions, extracted prices from e-commerce sites, and generated recommendations based on what it found. This was an advantage for stores with strong SEO and detailed product pages. After the update, the platform checks first whether a product exists in a directly connected catalog. If it does, it uses the structured data from the feed. If not, it might still find it through web search. But „might” is not a business strategy.
The effect was swift. Of 687 stores monitored by Profound, 450 lost at least 33% of their visibility in the first three days after the update. Only 67 gained visibility. Concentration intensified: the top 10 stores went from 22.5% to 41.8% of total product references, while unique merchants cited dropped from 13,524 to 10,607.
Fewer stores appear, but the ones that do dominate completely. Not because they have better content or stronger SEO, but for one reason: they have a product feed connected to the platform.
Shopify is already in. WooCommerce isn't.
Roughly 35% of feed-based recommendations come from Shopify Catalog, which is pre-integrated with ChatGPT. If you sell on Shopify, your products are already visible in Shopping conversations with zero extra effort. The same goes for Etsy, which has natively connected catalogs.
For other e-commerce platforms, the situation is different. WooCommerce, Magento, custom platforms, regional marketplaces: none of them have direct integration with ChatGPT Shopping right now. Salesforce and Stripe appear as feed providers with platform access, but the connection process is neither simple nor standardized.
What matters most is the model Profound built from this data. A statistical model that explains 83% of visibility variation among stores based on a single factor: whether the store has a connected product feed or not. Not content quality, not site speed, not review count. The feed.
This marks a structural shift. For 15 years, e-commerce optimization has been about content, user experience, and technical SEO. Now, before any of those, comes a binary question: are you in the platform's database, or aren't you?
It's a pattern we've seen before. Google Shopping worked on the same principle a decade ago: stores with product feeds dominated Shopping results, those without feeds simply didn't exist in that channel, regardless of their organic SEO. The difference is that Google Shopping was a separate tab many users ignored. ChatGPT Shopping is integrated into the main conversation. No tab to click, no context switch. You ask what you want to buy and get recommendations right there.
OpenAI has announced a self-serve platform for feed connections, expected toward the end of 2026. But in the meantime, stores already connected are gaining ground every day. And since ChatGPT Shopping doesn't show dozens of results like Google, but 3-5 recommendations per conversation, the available spots are few and filling fast.
What brands need to do right now
ChatGPT Shopping is not yet a major traffic source for most markets outside the US. But the trajectory is clear, and waiting for volume before acting is exactly how brands lost ground in Google Shopping a decade ago. Conversational advertising in ChatGPT was the first signal. Shopping is the second. The direction is singular: AI is becoming a purchase channel, not just an information channel.
Three concrete actions are worth taking now, not when the self-serve platform goes live.
First: check whether your e-commerce platform has or is planning integration with AI platform feeds. Not just Google Merchant Center. Ask your platform provider directly. If the answer is „we'll evaluate it,” treat that as a „no” and start actively looking for alternatives.
Second: invest in the quality of the product feed you already have. Google Merchant Center taught you the fundamentals, and they still apply: descriptive titles, correct categories, quality images, real-time pricing, live availability data. Profound anticipates the next phase of competition won't be feed access (which will become universal), but data granularity: more detailed fields, more specific attributes, more complete product information. A weak feed now means a structural disadvantage later. Think of it as the difference between having a store on a busy street versus a side alley: location matters, but so does what people see in the window.
Third: don't wait. Stores with connected feeds are gaining ground every day. Search behavior is already shifting toward AI, and purchasing will follow. Marketing teams that ignore AI feeds now will discover they've lost not just a channel, but a buying habit their customers have already formed.
Product feeds are no longer a technical e-commerce detail. They're the visibility infrastructure for the next generation of shopping platforms. Ignoring them doesn't cost you ranking. It costs you presence in the conversation that leads to sale.
