How to track ChatGPT and AI traffic in Google Analytics 4
GA4's new AI Assistant channel misses some assistants and app clicks land in Direct. The custom channel group, the regex, and what GA4 still can't see.
AI shopping is here - ChatGPT recommends products, compares them and even checks out. How ecommerce brands get surfaced: product data, reviews, structured feeds, and why it's relevance, not ad spend.
For most of the web's history, ecommerce visibility meant one thing: rank on Google, win the click, close the sale. AI assistants are quietly rewriting that. People now ask ChatGPT "what's the best running shoe for flat feet under $120?" and get a short list of named products - sometimes with a checkout button attached. If your product isn't in that answer, the click you used to compete for never happens. Here's how ecommerce brands get surfaced.
The most important thing to understand about AI shopping is what it isn't. When ChatGPT shows products, the results are organic and unsponsored - OpenAI has been explicit that paid checkout participation doesn't influence product ranking. Engines weigh relevance, availability, price and quality. That's a genuinely different game from Amazon's pay-to-play shelf: you can't buy your way to the top, which means a smaller brand with better data and reviews can out-rank a bigger one that spends more.
An engine can only recommend a product it can read clearly. That makes your structured product data - the same feeds and markup you built for Google Shopping - the primary lever:
Recommendations need justification, and reviews are where engines get it. Volume, recency and sentiment all feed how - and whether - a product gets named. This is the same reason review sites are an AI citation engine: the model is looking for third-party proof, not your own claims. A steady flow of genuine reviews is product-visibility infrastructure, not a support vanity metric.
Our own citation data shows engines lean heavily on review publishers, community threads and aggregators. For ecommerce that means "best X" roundups, YouTube reviews, Reddit recommendations and comparison sites are often what the engine reads before it answers. Getting your product into that third-party layer is as important as your own PDP.
ChatGPT's shopping features - recommendations, comparison and in-chat checkout - are the most developed today, and worth a dedicated play while they're still early. We broke that down in ChatGPT Shopping is the new shelf. And looking further out, AI agents that buy on a shopper's behalf raise the stakes on exactly the same fundamentals.
Curious whether AI already recommends your products - or your competitors'? A free Zene audit shows what the engines say when shoppers ask about your category.
Make your product data clean, complete and crawlable - accurate titles, prices, availability and specs, marked up with Product schema - and earn genuine reviews. ChatGPT's product results are organic and ranked on relevance, availability, price and quality, not on ad spend, so the levers are the same things that make a good product detail page: clarity, structure and social proof.
Not in the ad sense. OpenAI has been explicit that paid checkout participation doesn't influence product ranking, and results aren't sponsored. Merchants may pay a small fee on completed purchases, but that doesn't affect placement. This is closer to a relevance-ranked marketplace than to pay-to-play search ads.
More than ever. The structured product feeds, schema, reviews and clean detail pages you built for Google Shopping are exactly what AI shopping engines read. The work transfers directly - AI shopping is largely an extra consumer of the same well-structured commerce data.
Put this guide into practice - get your free visibility score in minutes.
