Social Rocket
AI assistants are becoming new gatekeepers in online shopping: a recent analysis by Jellyfish shows that ChatGPT, Google AI Mode and Amazon Alexa for Shopping select products and merchants very differently. For brands, this means that classic search visibility is no longer the only thing that matters. It is also about which shops and sources their products appear in for AI systems at all.
ChatGPT, Google and Amazon see different merchants
Digiday reported on 1 October 2026 about an analysis using Jellyfish's Share-of-Model technology. In a toy query, Alexa for Shopping showed products from 177 brands, almost exclusively from Amazon's own store. ChatGPT referred to 24 merchants for the same type of search, and Google AI Mode to 37.
This finding is relevant because AI shopping is not based on a uniform index. Which brand gets recommended also depends on which merchants, product pages and sources the respective system takes into account.
This shift fits with our overview of GEO as a new publisher and media business and with Google's Preferred Sources in AI Overviews and AI Mode.
Merchant presence becomes part of AI visibility
Jellyfish describes its Share-of-Model platform as a tool that lets brands measure how AI models perceive and recommend them. The company argues that LLMs and shopping agents are increasingly taking over research, comparison and product selection themselves.
In practical terms, this means product data must be clean not only on the brand's own website. Availability, specifications, merchant coverage, reviews and independent content can all influence whether a product ends up in an AI recommendation at all.
A similar effect is already visible with Meta Muse and agentic shopping via WhatsApp. The role of platforms is also shifting in retail media, as our look at Walmart Connect and offsite advertising shows.
AI traffic is small, but more valuable
Adobe data, which Reuters reported on in June, also shows a shift: in May, users who arrived at US retail websites via ChatGPT, Gemini or other AI systems generated 53 percent more revenue per visit than visitors from other sources. According to Adobe, the conversion rate was also higher.
These figures do not mean that AI traffic has already replaced classic search. They do show, however, that users who receive an AI recommendation often arrive at merchant sites with a higher purchase intent.
What marketers should check now
For e-commerce and social commerce teams, an additional visibility check is therefore worthwhile: which products are mentioned by ChatGPT, Google AI Mode, Gemini or other assistants? Which merchants and sources appear in the process? And which products from your own brand are missing entirely?
GEO and agentic commerce are thus converging. It is not only about being mentioned in an AI answer, but about appearing in the right purchase context with correct product data and trustworthy sources.