By Steve Merrill · October 26, 2026
Where do Perplexity, ChatGPT, and Google AI actually get the products they recommend?
They pull from three different places. Perplexity ranks live web pages and reviews, ChatGPT reads structured product feeds inside its own shopping flow, and Google AI Overviews pull from the Shopping Graph fed by Merchant Center. Same question, three different pipes. If you tune for one and ignore the other two, you show up in a third of the answers.
I've watched merchants chase "AI visibility" like it's one switch. It isn't.
These three engines recommend products in completely different ways. Let me break down how each one works and what you actually do about it.
How does Perplexity pick which products to recommend?
Perplexity recommends products by ranking live web pages, then citing the best ones with a footnote. It reads your product page, reviews, comparison articles, and Reddit threads, then summarizes an answer with links back to sources.
Perplexity is a search engine with an answer layer on top. Its Sonar models run real web searches and cite what they find, which the company documents in its own developer docs.
So the game here is old-school in a new coat. Strong product pages. Real reviews. Mentions on pages Perplexity already trusts. If a comparison article ranks and names your product, you get pulled into the answer.
What to do for Perplexity:
- Write product pages that answer buyer questions in plain text, not just specs.
- Get reviews on your page and on third-party sites. Perplexity loves consensus.
- Earn mentions in "best X for Y" roundups. Those get cited constantly.
- Keep your page fast and crawlable. No answer engine fights your JavaScript for you.
How does ChatGPT surface products when someone asks for a recommendation?
ChatGPT surfaces products through its built-in shopping experience, which reads structured product data from merchants and partners and shows product cards with price, image, and reviews. It leans on feeds, not just a web crawl.
This one changed the math. OpenAI rolled out shopping results that pull product metadata like price, availability, and reviews, and they've said these results are organic, not ads. You can read their own writeup on buying inside ChatGPT.
Here's the bottom line. ChatGPT wants clean, structured product data it can trust. Missing a price or a GTIN and you look broken next to a competitor who filled every field.
What to do for ChatGPT:
- Fill out your full product feed. Title, price, availability, GTIN, images, condition.
- Keep availability and price accurate in real time. Stale data gets you dropped.
- Pile up genuine reviews. The product card shows them, and they move clicks.
- Use specific, descriptive titles. "Merino wool base layer, men's, lightweight" beats "Shirt 2."
A clean feed is doing double duty now. Same data that powers Google Shopping is starting to power ChatGPT.
How does Google AI handle product recommendations?
Google AI Overviews recommend products using the Shopping Graph, a massive live dataset of products, prices, and reviews fed largely by Google Merchant Center. When someone asks a shopping question, the Overview pulls from that graph and your organic pages.
Google says its Shopping Graph holds over 50 billion product listings, and it refreshes billions of those every hour. That scale is the whole point. Google covers the AI shopping features on its own blog.
For Shopify merchants this is the most direct path, because Shopify already syncs products to Google through the Google & YouTube channel. Your feed is probably half-built already. The question is whether it's any good.
What to do for Google AI:
- Connect Shopify to Google Merchant Center and fix every feed error. Don't ignore the warnings tab.
- Add product structured data (schema.org/Product) to your pages so organic results reinforce the feed.
- Keep GTINs, brand, and category fields accurate. Google matches products across sellers with them.
- Collect Google-visible reviews through Product Ratings. They show up right in the Overview.
So which engine should a Shopify store tune for first?
Start with Google, because it has the widest reach and you're half set up already. Then build one strong product feed that also feeds ChatGPT. Then earn the reviews and roundup mentions that get you into Perplexity. One clean feed plus strong pages covers most of all three.
Notice the overlap. A complete, accurate product feed helps you in Google AI and ChatGPT. Strong pages with real reviews help you in Perplexity and Google. The work stacks.
Here's what I mean. I audited a home goods store last quarter that showed up in zero AI answers. Their feed had 60% of products missing a GTIN and half their titles said "Default Title." We fixed the feed, added Product schema, and pushed for reviews. Six weeks later they were getting cited in both Google Overviews and ChatGPT cards. Same products. Better data.
The data doesn't lie. These engines reward the merchant who fills every field and answers every buyer question. Boring work. It wins.
What do the three engines have in common?
All three reward structured data, real reviews, and pages that answer questions directly. None of them will fix your messy feed or your thin product descriptions for you. The inputs are the same even when the pipes differ.
That's the quiet truth under all the AI hype. The fundamentals got more important, not less. Clean data. Honest reviews. Clear writing. The engines just raised the stakes on getting them right.
Frequently asked questions
Do Perplexity, ChatGPT, and Google AI pull product data the same way?
No. Perplexity ranks live web pages and reviews. ChatGPT pulls from merchant and partner product feeds inside its shopping experience. Google AI Overviews lean on the Shopping Graph fed by Merchant Center. Three different pipes, three different fixes.
Which AI engine sends the most shopping traffic to Shopify stores right now?
Google AI Overviews reach the widest audience because they sit on top of normal search. ChatGPT drives the most intent-heavy shopping sessions through its product cards. Perplexity is smaller but converts well for research-stage buyers.
Do I need a Google Merchant Center feed to show up in AI shopping answers?
For Google AI Overviews and ChatGPT product cards, a clean structured feed is the price of entry. Perplexity can cite you from a well-built product page alone, but a feed still helps everywhere.
How fast do changes to my feed or pages show up in AI answers?
Feed changes in Google Merchant Center can refresh within a day. Perplexity picks up page edits as it recrawls, often within a week. ChatGPT updates on a cadence tied to feed syncs and index refreshes.
Ready to show up in all three?
We build the feed, the schema, and the pages that get Shopify products recommended across Perplexity, ChatGPT, and Google AI. No hype. Just the boring work that wins.
