GEO for online stores: how ChatGPT, Google and Perplexity pick products

20 October 2026 · 5 min read

Our earlier guide to AI shopping answers covered the basics: clean product data, strong category pages and buying guides. This one goes a step further. It looks at what ChatGPT, Google and Perplexity say, in their own documentation, about how they choose which products to show, and turns that into a short list of fixes for a store owner.

ChatGPT: product metadata and feeds

OpenAI says ChatGPT selects products using structured metadata, such as price and product description, from the merchant or from third-party providers, along with other third-party content.
When several stores sell the same item, OpenAI says it ranks them on factors like availability, price, quality and whether the store is the maker or main seller.
Shopify stores are already connected through Shopify Catalog, and OpenAI says they need to do nothing extra. Other stores can apply to send OpenAI a product feed directly.
OpenAI says product results are not ads and are not influenced by its partnerships. You cannot pay to appear.

Google: Merchant Center and Product markup

Google takes product data in two ways: Product structured data on your pages, and a product feed in Google Merchant Center with free listings turned on. You can use either, or both. Its merchant listing markup can also carry shipping costs, return policy and apparel sizes. Merchant Center help now talks about how products are discovered in AI Mode, AI Overviews and the Gemini app, so the feed is part of how products reach those features, not only Shopping results.

Perplexity: relevance and ratings

Perplexity says it ranks product listings much as it ranks sources, on authority and relevance, and that listings are not sponsored. Its help pages name relevance and ratings as what decides which products are shown.

What this means for a store

All three point the same way. The assistant decides from the facts it can read about a product, not from how the page looks. A product with a missing price, unclear stock or no brand is easy to skip, however good the photos are.

Fix these first

Product markup on every product page with name, image, description, brand, price, currency and availability, matching what the page shows.
A SKU, and a GTIN where your products have barcodes, so shopping systems can match your listing to the right product.
Shipping and returns in your markup or Merchant Center settings. They appear next to products and help shoppers compare stores.
A Google Merchant Center feed with free listings turned on and no disapproved items.
If you are not on Shopify, read OpenAI's merchant page and consider applying to send a direct product feed.
A robots.txt and firewall that let OAI-SearchBot, PerplexityBot and Googlebot reach product pages, with product details in the HTML rather than only after scripts run.
Genuine reviews, marked up honestly. Never mark up reviews that customers did not write.

How to tell if it is working

Ask the assistants the questions your buyers ask, such as "best running shoes for flat feet under $100", and note whether your products appear and whether the price and stock are right. Repeat the same questions every month. Answers change from one run to the next, so judge the pattern over several checks, not a single result. Our Store Report checks a store's product markup, crawler access and technical basics, and ranks what to fix first.

As with all AI search work, nobody can guarantee a product will be recommended. What a store controls is whether its product facts are complete, correct and readable by machines.

Sources

Free tools: AI crawler checker · llms.txt generator · robots.txt AI rules generator