To optimize product pages for ChatGPT and AI search, publish specs, prices and availability as plain text, answer the comparison questions shoppers ask, add Product markup that matches the page, earn reviews on sites AI engines read, and send a product feed to OpenAI and Google Merchant Center. Engines pick products from structured, consistent data.
Key takeaways🔗
- OpenAI says ChatGPT picks products using structured metadata from first-party and third-party providers (such as price and description), other third-party content, and the model's own knowledge.
-
ChatGPT ranks merchants for a product on factors like availability, price, quality, and whether you are the maker or primary seller.
-
Specs, sizes, materials and compatibility should be plain HTML text, not images or tabs loaded by JavaScript.
-
Product schema and a Merchant Center feed help Google. OpenAI has its own product feed spec, and Shopify stores are already connected through Shopify Catalog.
-
Review summaries in ChatGPT come from reviews on public websites, so reviews off your site matter as much as the ones on it.
How does ChatGPT choose which products to show?🔗
ChatGPT chooses products by matching the shopper's intent against product data it gets from merchants, data providers and the open web. OpenAI's help article on shopping with ChatGPT search lists three inputs: "structured metadata from first-party and third-party providers (e.g., price, product description) and other third-party content," model responses generated before any new search, and OpenAI's safety standards and product policies.
The same article says a few more things every store owner should know:
- Product results are not ads. OpenAI says they "are selected independently by ChatGPT and are not ads, nor influenced by any OpenAI partnerships." Ads are shown separately.
-
Merchants are ranked "based on factors like availability, price, quality, and whether they are the maker or primary seller of that item."
-
Titles and descriptions may be rewritten. ChatGPT "may generate simplified product titles and descriptions based on information it receives from third-party providers."
-
Review summaries come from the web. They are "based on reviews from public websites" and are not verified by OpenAI.
-
Prices can lag. When merchants update pricing or shipping, "there may be some delay before it is reflected."
Put together, that means ChatGPT rewards products whose facts are easy to extract and consistent everywhere they appear. Our broader guide to ecommerce AI search visibility covers the store-level side; this post is about the product page itself.
What should a product page include for AI search?🔗
A product page that AI engines can use states every buying fact in plain text, close to the top, in words a shopper would use in a question. If a shopper might ask it, the answer should be on the page as text.
| Element | Weak for AI | Strong for AI |
|---|---|---|
| Title | "The Voyager" | "Voyager 28L waterproof hiking backpack" |
| Specs | Spec sheet as an image or PDF | HTML list: weight, dimensions, materials, capacity |
| Use case | "Built for adventure" | "Fits a 16-inch laptop. Carry-on size for most US airlines." (if true) |
| Price and stock | Loaded by script after page load | In the server-rendered HTML, with currency |
| Variants | Only visible after clicking a swatch | Every size and color listed in text |
| Policies | Linked in the footer | Shipping time and return window stated on the page |
Server rendering matters more than most stores realize. Vercel's study of AI crawler traffic, The rise of the AI crawler (December 2024), found that "none of the major AI crawlers currently render JavaScript," including the crawlers from OpenAI, Anthropic and Perplexity. Specs inside a JavaScript tab or a review widget that loads after the page may simply not exist for those crawlers. To check, use "View page source" and search for a spec like the weight. If it is not in the source, move it.
Write the description for the questions shoppers actually ask an AI assistant: "Is it waterproof?", "Will it fit a 16-inch laptop?", "Is it good for a beginner?", "How does it compare to [competitor model]?" A short Q&A block under the description is a natural place for these.
Why does comparison content help products get recommended?🔗
Comparison content helps because a large share of shopping prompts are comparisons ("X vs Y," "best X under $100," "X alternatives") and the engine needs a source that has already done the comparing. OpenAI's help article notes that shoppers can ask ChatGPT "to compare products, explain trade-offs, or suggest alternatives."
You can supply that source yourself:
- On the product page, add a short "How it compares" section with a table against your own other models, so the engine can tell your 28L from your 40L.
-
On separate pages, publish honest comparisons against the competitor products shoppers name most often. Include the cases where the other product is the better pick. One-sided pages are easy for both shoppers and models to discount.
-
On buying guides, answer "best X for Y" questions with clear criteria and dated prices.
Then find out whether it is working. A competitor AI visibility analysis shows which products the engines name instead of yours, and which sources they cite to do it.
What Product schema do you need?🔗
You need Product markup with a nested Offer that carries the same name, price, currency and availability shown on the page, plus ratings only if they come from real reviews. Google's product structured data docs describe two uses: product snippets for pages where shoppers cannot buy, and merchant listings for pages "where customers can purchase products from you," which add options such as shipping details and return policy.
The fields that matter most for shopping questions:
name,description,brand,image, and an identifier such asgtinorsku
-
offerswithprice,priceCurrencyandavailability -
aggregateRatingandreview, only when they reflect real reviews shown on the page -
shippingDetailsandhasMerchantReturnPolicyfor shipping and returns
Keep two warnings from Google's merchant listing docs in mind. First, "dynamically-generated markup can make Shopping crawls less frequent and less reliable, which can be an issue for fast-changing content like product availability and price." Generate it on the server. Second, Google recommends defining shipping and return policies once at the organization level where they apply store-wide.
Be realistic about what markup does for AI answers. Google's AI features guide states that "structured data isn't required for generative AI search, and there's no special schema.org markup you need to add." Markup makes your facts unambiguous. It does not make a weak product page persuasive. For full JSON-LD examples, see our schema markup guide.
How much do reviews matter for AI product recommendations?🔗
Reviews matter a lot, and the reviews on other websites matter at least as much as the ones on your own page. OpenAI says ChatGPT's product review summaries "are based on reviews from public websites and are intended to highlight common user likes and dislikes about a product," and that labels like "Budget-friendly" can come from what reviewers frequently mention.
What to do about it:
- Show reviews in the HTML. If your review widget loads by script, crawlers that do not run JavaScript see a page with no reviews.
-
Ask for specific reviews. A review that says "kept my laptop dry in a two-hour rainstorm" gives an engine a usable fact. "Great product!" does not.
-
Be present where your category is reviewed. That might be Amazon, specialist review sites, YouTube, or Reddit threads. Our post on Reddit citations in AI answers explains why community threads show up so often.
-
Answer complaints in public. A clear, polite reply to a common complaint becomes part of the record the engine summarizes.
Do not mark up reviews you do not show, and never write fake ones. Neither holds up when a model reads the wider web and finds something different.
Should you submit a product feed to ChatGPT and Google?🔗
Yes, if you sell physical products, because feeds are the most direct way to give AI shopping features your current price and stock. Pages get crawled now and then; feeds can be updated when something changes.
| Channel | What to do | Source |
|---|---|---|
| ChatGPT (Shopify stores) | Nothing extra. OpenAI says Shopify product data is already integrated through Shopify Catalog. | OpenAI Help Center |
| ChatGPT (other stores) | Read OpenAI's product feed docs and apply for direct feed access. | OpenAI product feed spec |
| Google AI Overviews and AI Mode | Keep a Google Merchant Center feed and on-page structured data. | Google product docs |
OpenAI's feed spec has nine required fields: item_id, title, description, url, brand, seller_name, image_url, availability and price. It also has optional review_count and star_rating fields, and tells merchants to update the feed when a sale starts or ends and when an item sells out. Google, for its part, says that "providing both structured data on web pages and a Merchant Center feed maximizes your eligibility to experiences," and its AI features guide names Merchant Center as a way to help products appear in AI responses.
The rule for both: the feed, the page and the markup should say the same thing. Since OpenAI ranks merchants partly on price and availability, and highlights the lowest price it knows about with a "Best price" label, a stale feed price works against you.
How do you know if your product pages are showing up?🔗
You find out by asking the engines the questions your shoppers ask, such as "best [category] under $[price]," "[your product] vs [competitor]," and "is [your product] worth it," and recording which products and sources come back. Repeat the same prompts over time, since AI answers change from one run to the next.
I built TrendlyAI to do that on a schedule: it runs your prompts weekly on ChatGPT, Perplexity, Claude and Google AI Overviews, shows when a competitor is named and you are not, and drafts fixes such as a product page rewrite or a comparison page. It does not track Gemini, Copilot or AI Mode, so if those matter most to you, check them by hand too. For a quick read today, the free AI visibility check runs 5 prompts on ChatGPT and Perplexity in about 30 seconds, no signup. For the full playbook, see the GEO hub, and for a step-by-step on tracking, see how to track brand mentions in ChatGPT, Perplexity and Google.
Frequently asked questions🔗
Are ChatGPT product results paid ads?🔗
No. OpenAI says product results in ChatGPT are selected independently and are not ads, nor influenced by any OpenAI partnerships. Ads in ChatGPT are shown separately from product results.
Do I need Product schema to show up in ChatGPT?🔗
Not strictly. OpenAI says ChatGPT uses structured metadata from merchants and third-party providers, which includes feeds as well as page data. Product schema still makes your price, stock and ratings unambiguous, and Google uses it for merchant listings.
How do I get my products into ChatGPT shopping?🔗
If you sell on Shopify, OpenAI says your product data is already integrated through Shopify Catalog. Other merchants can read OpenAI's product feed documentation and apply for direct feed access. Clear, crawlable product pages help in both cases.
Why does ChatGPT show the wrong price for my product?🔗
Usually because its data is stale or comes from a third-party listing. OpenAI notes there can be a delay before pricing updates show up. Keep your page, markup and feed in sync, and update the feed when prices change.
Does ChatGPT read the reviews on my product page?🔗
ChatGPT's review summaries are based on reviews from public websites, which can include yours if they are in the page HTML. Reviews on other sites in your category also feed those summaries, so earn reviews beyond your own store.