Short answer: Product page SEO for AI means putting the facts a shopper needs to choose in plain text on the page: what the product is, who it's for, the key specs, the price, stock, delivery and returns. Answer the questions buyers ask, show real reviews, and add Product structured data that matches the page. Then send the same data to Google Merchant Center and, where you can, to ChatGPT.
What is product page SEO?
Product page SEO is making the page for one product easy for search engines to understand and show to shoppers. For years that meant ranking in Google's results. Now the same page also feeds the shopping answers in ChatGPT and Google's AI Mode, which pick a handful of products and say why.
Google says there are no extra requirements for its AI features, and "no special schema.org structured data that you need to add". Its checklist for those features does include keeping your Merchant Center information up to date. For an online store, that means your product data.
What does ChatGPT use when it recommends a product?
OpenAI's help page on shopping in ChatGPT says product results "are not ads". ChatGPT uses structured product data from stores and from third-party providers, such as price and description, and summaries of reviews from public websites. It ranks merchants on factors that include "availability, price, quality", and it looks at whether a store is the maker or primary seller of the item.
What we saw in 3 real shopping searches
On 16 October 2026 we asked ChatGPT 3 shopping questions: the best cast iron skillet for a beginner, a chew-proof bed for a large dog, and running shoes for flat feet. It recommended 11 products.
- Every pick came with checkable facts. Almost every pick had a price, and each came with the details that mattered, like size, material or the type of support. The dog beds also showed a star rating and review count. The top skillet came with "Downside: ... the pan is heavy."
- The reasons came from articles. ChatGPT cited 15 review, buying-guide and how-to articles across the 3 answers.
- Product data supplied the buy links. 3 of the store links carried tags showing they came from a product feed. One went to a small brand's own store: K9 Ballistics, the top dog bed, shown as "Sold by K9 Ballistics" at $135 with 4.8 stars from 2,950 reviews.
- Some brands got only a home page. All 5 running shoes linked to the brand's home page, not the shoe. That's a poor place to land for someone ready to buy one shoe.
So a product page has two jobs. It has to state the facts in a form a machine can read and compare. And it has to be a good page to land on when ChatGPT sends a shopper to buy.
How to structure a product page so AI can recommend it
1. Name the product the way people search for it
Use the brand, the product type, the model and the one attribute that matters most: size, capacity or material. ChatGPT's own product cards in our test used names like "Lodge Classic 10.25-inch Skillet". "Ridgeway 30L waterproof hiking daypack" tells a machine what it is. "The Ridgeway" doesn't.
2. Open with a short summary in plain text
Write 2 to 3 sentences near the top: what the product is, who it's for, and the fact that sets it apart. Put the price, stock and delivery time next to the Buy button. Keep all of it in text, not in an image. Google asks for important content "in textual form".
3. Put the specs in a table
Dimensions, weight, material, capacity, compatibility, what's in the box and the warranty. Give units every time. Specs are a list of attributes, so this is the place on a product page where a table is the right format.
4. Say who it's for, and who it isn't for
Every pick in our test came with a line on who it suits or a downside. Write those lines yourself: "Best for day hikes and wet commutes. Too small for an overnight trip." An AI tool can repeat a sentence like that, and a shopper who reads it is less likely to send the product back. Don't hide the limits.
5. Answer buyer questions under question headings
Cover fit and sizing, care, compatibility and how long delivery takes. Find the questions in your customer emails, your return reasons and the questions people ask about similar products on marketplaces. As in part 1, answer each one in the first sentence under the heading.
6. State price, stock, shipping and returns on the page
OpenAI names availability and price among the ways it ranks merchants. Show whether the product is in stock, the delivery time in days, the return window and who pays return shipping. Don't leave the terms only on a separate policy page.
7. Show real reviews
Show reviews from your buyers on the product page, with the number of reviews. Only mark up reviews that are on the page. Google's rule is "Don't mark up content that is not visible to readers of the page."
8. Write your own description
Keep the manufacturer's specs, since changing them risks errors. But write the description yourself. If you paste the manufacturer's text, your page says what every other seller's page says. Add what only you know: how it compares with the next model, what customers ask before buying, and what they return it for.
9. Add Product structured data, and send the same facts to shopping feeds
Mark up the page as a Product with an Offer. For a page where people can buy, Google requires the name, image, price and currency, and recommends the brand, SKU, GTIN, availability, rating and reviews, and shipping and return details. Google suggests one store-wide shipping and return policy in your Organization markup, with product-level policies only for exceptions.
Then send the same facts as a feed. Google says providing both structured data and a Merchant Center feed maximizes your eligibility for its shopping experiences. For ChatGPT, OpenAI says Shopify stores are already connected through Shopify Catalog, and other stores can apply to send a direct product feed. Your store platform or SEO plugin may already output some of the markup, so check what your page outputs with Google's Rich Results Test.
Before and after: one product page rewritten
Here's the top of a typical product page from a made-up outdoor store, then the same page restructured.
| Before | After |
|---|---|
| The Ridgeway Pack Adventure awaits! Our best-selling pack is built for the trail and beyond. Premium materials, smart design and all-day comfort make it the perfect companion for whatever you do. Order yours today! |
Ridgeway 30L waterproof hiking daypack A 30-litre daypack for hikes of up to a full day, with a waterproof roll-top and a padded hip belt that takes weight off your shoulders. It weighs 1.1 kg (2.4 lb). Best for day hikes and wet commutes. Too small for an overnight trip, so see the 45L. $129. In stock. Ships in 1 business day, US delivery in 3 to 5 days. Free returns within 30 days. |
Under the summary, the specs go in their own table:
| Spec | Ridgeway 30L |
|---|---|
| Capacity | 30 litres |
| Weight | 1.1 kg (2.4 lb) |
| Fits torso length | 16 to 20 inches |
| Material | Waterproof coated nylon, taped seams |
| Laptop sleeve | Fits up to a 15-inch laptop |
| Warranty | 5 years |
The "after" version answers what a shopper would ask ChatGPT. How big is it? How heavy? Is it waterproof? Who's it for? When will it arrive? The "before" version doesn't answer one of them.
A product page outline you can copy
| Part | What goes in it |
|---|---|
| Product name (H1) | Brand, product type, model and the key attribute |
| Summary | 2 to 3 sentences: what it is, who it's for, what sets it apart |
| Price and stock | In text next to the Buy button, with the delivery time |
| Spec table | Dimensions, weight, material, compatibility and warranty, with units |
| Best for, not for | One line each |
| Buyer questions | 3 to 6 question headings, each answered in the first sentence |
| Reviews | Real reviews from buyers, with the count |
| Shipping and returns | Delivery time in days, the return window, who pays return shipping |
| Structured data and feeds | Product and Offer markup that matches the page, plus a Merchant Center feed |
What to avoid
- Specs only in images or PDFs. Text in a picture or a download is harder to read than text on the page.
- Slogans instead of facts. "Premium quality" and "the perfect companion" give an AI tool nothing to compare.
- Markup that disagrees with the page. If the markup says in stock and the page says sold out, one of them is wrong, and Google's rules say markup must describe what's visible.
- Shipping and returns hidden behind a link. Keep the full policy page, but put the key terms on the product page as well.
Does ChatGPT recommend your store?
Our free AI visibility check tests whether AI tools can read your site, then emails you a report on what ChatGPT and Google's AI say when asked for a business like yours.
Next in this series: how to structure a category page so AI sends shoppers to it. Part 1 is how to structure a blog post for AI. For the bigger picture, read how to get recommended by ChatGPT, or browse all our guides.
Frequently asked questions
What is product page SEO?
Product page SEO is making the page for one product easy for search engines and AI tools to find, understand and recommend. It covers both what shoppers read on the page and the Product structured data behind it.
Does ChatGPT use my product pages to recommend products?
ChatGPT uses structured product data from stores and third-party providers, such as price and description, plus summaries of reviews from public websites, according to OpenAI. Shopify stores are connected through Shopify Catalog. Other stores can apply to send OpenAI a direct product feed.
Do I need Product schema to appear in Google's AI Mode?
No. Google says its AI features need no special structured data. Product markup and a Merchant Center feed still matter for shopping results, and Google says providing both maximizes your eligibility for its shopping experiences.
How long should a product description be?
There's no set length. Cover what a buyer needs to decide: what it is, who it's for, the specs that matter, and the questions people ask. A short description full of facts gives an AI tool more to use than a long one full of adjectives.
Should I use the manufacturer's product description?
Use the manufacturer's specs, but write the description yourself. If every store uses the same text, neither a shopper nor an AI tool has a reason to pick yours. Add what you know from selling it: comparisons, common questions and fit.