For a quarter of a century, the playbook for selling products online remained largely unchanged: optimize your website, build authority, and hope Google’s crawlers rank you on page one. But as generative artificial intelligence begins to reshape how consumers search, that playbook is being rewritten. A quiet but massive shift is underway inside OpenAI’s ChatGPT, one that suggests the future of e-commerce discovery may rely less on traditional search engine optimization (SEO) and far more on structured data feeds.
The Great Feed Migration
According to data from e-commerce analytics firm Profound, ChatGPT’s shopping recommendations have undergone a dramatic structural realignment. Profound, which tracks daily prompt logs to analyze product recommendations, observed that the share of “feed-integrated” product recommendations in ChatGPT rose from a mere 8.26% to a staggering 61.54% on July 10. This analysis was based on 1,757,723 tracked prompt runs throughout the month of July.
This means ChatGPT is increasingly bypassing traditional web-crawled content—such as blog reviews, listicles, and organic site pages—in favor of structured product feeds. These feeds, typically provided directly by merchants, contain clean, standardized data about pricing, availability, and product specifications.
The impact of this shift on brand visibility was immediate and highly volatile. From the period of July 7–9 to July 10–12, 450 of the 687 brands tracked by Profound experienced at least a one-third decrease in their ChatGPT Shopping visibility. Conversely, 67 brands saw their visibility jump by at least that same margin. This sudden redistribution of digital real estate highlights the risks of relying solely on legacy SEO strategies in an AI-first market.
Getting In Before Getting Ranked
This trend points toward a broader theme emerging in the search landscape: the necessity of “getting in” before you can even hope to rank. Traditionally, search engines crawled the open web and decided how to rank what they found. Today, platforms are increasingly demanding that businesses hand over structured data upfront to even be considered for display.
We are seeing this play out beyond OpenAI. Google’s recent redesign of commercial search results in the European Economic Area (EEA) features specialized units for comparison sites and direct suppliers. To appear in the aggregator units, comparison sites must actively feed data to Google via APIs or structured feeds. Similarly, ChatGPT’s heavy reliance on merchant feeds suggests that if a brand does not have a direct data pipeline into these AI systems, it may become entirely invisible to consumers who use AI as their primary shopping assistant.
A Very Large Online Search Engine
The stakes for mastering this new pipeline are incredibly high. ChatGPT is no longer just a niche conversational novelty; it has matured into a major infrastructural player. OpenAI recently reported that ChatGPT attracts an average of 159.1 million monthly search recipients in the European Union alone. This milestone officially designates ChatGPT as a “Very Large Online Search Engine” (VLOSE) under the EU’s Digital Services Act, placing it in the same regulatory tier as legacy search giants.
As ChatGPT cements its status as a primary search destination, its internal mechanics will dictate the fortunes of online retailers. For brands, the message is clear: the technical management of product feeds—once a back-office task relegated to IT or Google Merchant Center managers—has suddenly become a core pillar of modern digital marketing. To win in the age of generative search, businesses must ensure their data is structured, accessible, and ready for the machines.