---
title: "Product schema and product pages that AI answers can use"
url: https://marketingsignals.com/insights/product-pages-and-schema-for-ai-answers
type: Article
date: 2026-11-02
author: "Sam Forrester"
---

# Product schema and product pages that AI answers can use

Product schema, review schema and product page copy that Google rich results, Merchant Center and AI answers can read, and how to keep it all in sync.

## Key takeaways

- AI answers compare products on the facts they can read, so specifications, compatibility, use cases and policies belong on the page as text, not in images or PDFs.
- Retail product pages need merchant listing markup: Product with name and image, a complete Offer, shipping and returns details, identifiers and variants modelled with ProductGroup.
- Google says providing both structured data and a Merchant Center feed maximises eligibility for its shopping experiences, and the two must match what the shopper sees.
- Generate the page, the markup and the feed from one product record, and put Product markup in the initial HTML, because Google says JavaScript-generated markup makes Shopping crawls less reliable.
- Only mark up reviews that are visible on the page, and never aggregate ratings from other websites into your own markup.

When a shopper asks an AI assistant which product to buy, the answer is assembled from what machines can read about your products: the copy on the page, the product schema behind it and the data in your feed. If those three disagree, or the facts a shopper compares are missing, the answer either leaves you out or describes you wrongly.

Most retailers already have some structured data, usually whatever the platform or a plugin added years ago. This article sets out what we put on product pages and in the markup so that Google's rich results, Merchant Center and AI answers can use it, how we keep it consistent with the feed and how we test it.

## Why product data quality is a revenue question

Product pages are where the sale happens, so errors here cost directly. Google's [Merchant Center guidance on structured data](https://support.google.com/merchants/answer/6069143) says the markup must match the values shown to the customer, and that incorrect data on product landing pages breaks its structured data guidelines. Google also links price and availability mismatches to item disapprovals and account suspensions, and a disapproved product earns nothing from Shopping ads or free listings.

The upside is that one piece of work serves several channels. Google's [introduction to product structured data](https://developers.google.com/search/docs/appearance/structured-data/product) says that providing both structured data on web pages and a Merchant Center feed maximises your eligibility for its experiences. Its guidance for AI Overviews and AI Mode asks for structured data that matches the visible text and Merchant Center information kept up to date. And a page that states its facts plainly is the version any AI assistant can quote, whether or not it reads the markup.

## What AI answers quote from a product page

Structured data describes the product; the copy explains it. AI answers compare products, so the facts that matter are the ones a shopper would compare, written in text on the page:

- Specifications: dimensions, weight, materials, capacity and power, in an HTML table rather than an image or a PDF.
- Compatibility and fit: what it works with, what it replaces and size guidance with real measurements.
- Use cases: who it is for, the situations it suits and who it is not for.
- Comparisons: how it differs from the nearest alternatives in your range, which helps an answer recommend the right one.
- Policies: delivery times, delivery costs and returns for this product, stated on the page.
- Answers to the questions your customer service team hears most often.

Google's guidance on [AI features and your website](https://developers.google.com/search/docs/appearance/ai-features) asks site owners to keep important content available in text form, and says no special schema is needed for AI Overviews or AI Mode. In practice, a system that reads a page in passages can use a sentence that states a fact; it cannot use a fact that exists only in an image. When we audit product pages for [AI citations](https://marketingsignals.com/glossary/ai-citations), missing comparison facts are the gap we find most.

## How we implement product schema

We implement [structured data](https://marketingsignals.com/glossary/structured-data-for-ai-search) in JSON-LD, the format Google's [general structured data guidelines](https://developers.google.com/search/docs/appearance/structured-data/sd-policies) recommend, generated by the platform from the same product record as the feed. The order of work:

1. Use the right type for the page. Google separates product snippets, for pages where people cannot buy the product (such as editorial reviews), from merchant listings, for pages where customers can buy from you. A retailer's product pages need [merchant listing markup](https://developers.google.com/search/docs/appearance/structured-data/merchant-listing): Product with a name, an image and a nested Offer, all of which are required.
2. Fill in the Offer completely. Google recommends availability, itemCondition, priceValidUntil, shippingDetails and hasMerchantReturnPolicy on the Offer, and properties such as brand, gtin, color and description on the Product. Price and availability drift most often, so they must come from the live system rather than a cached template.
3. Model variants properly. Where a product comes in sizes or colours, Google supports [ProductGroup](https://developers.google.com/search/docs/appearance/structured-data/product-variants) with hasVariant for each variant, variesBy for the attributes that change (such as size or colour) and productGroupID for the parent SKU. If variants are selected on a single page, there must be only one canonical URL for the group.
4. Set shipping and returns once, then override. Google accepts a standard [return policy](https://developers.google.com/search/docs/appearance/structured-data/return-policy) at organisation level, through hasMerchantReturnPolicy on Organization, with Offer-level policies for the exceptions. Shipping is described with schema.org's [OfferShippingDetails](https://schema.org/OfferShippingDetails), which covers shippingRate, shippingDestination and deliveryTime.
5. Mark up only the reviews you show. Google's [review snippet guidelines](https://developers.google.com/search/docs/appearance/structured-data/review-snippet) require review content to be readily available on the marked-up page, require at least one of ratingCount or reviewCount for an aggregate rating, and say not to aggregate reviews or ratings from other websites.
6. Put the markup in the initial HTML. Google recommends Product structured data in the initial HTML for best results, and warns that JavaScript-generated markup can make Shopping crawls less frequent and less reliable. On headless and JavaScript-heavy builds this is the first thing we check.

Where shipping or return policies are defined in more than one place, Google applies an order of precedence, strongest first: product-level feeds in Merchant Center, Content API for Shopping settings, settings in Merchant Center or Search Console, product-level markup, then organisation-level markup. We record which source is the master for each policy so nobody edits the weaker one and wonders why nothing changes.

Star ratings in Shopping ads and free listings come through a separate route. Merchant Center's [Product Ratings](https://support.google.com/merchants/answer/14549080) programme needs at least 50 reviews across all your products, submitted through a supported aggregator or directly, with the review data uploaded at least once a month.

## Keeping product schema in sync with the feed

The feed and the page drift apart when they come from different systems or update at different times. A sale price changes in the feed overnight while a cached page still shows the old price, or a product sells out and the markup still says InStock.

We fix this at the source. The page, the markup and the [Google Merchant Center](https://marketingsignals.com/glossary/google-merchant-center) feed should all read from one product record, and cached pages should be cleared when price or stock changes. Merchant Center can use valid markup for automatic item updates to price and availability, which Google says reduce the risk of disapprovals from mismatches, but it also says these updates are not a replacement for keeping feed data current.

Once the core data is clean, Merchant Center's conversational attributes are worth adding. Our guide to [Merchant Center feed optimisation](https://marketingsignals.com/insights/google-merchant-center-feed-optimisation) covers the feed side in detail.

## How to test and measure it

- Before release: run each template (standard product, variant, out of stock, on sale) through the Rich Results Test and check that every value matches what the page shows.
- After release: watch the Merchant listings and Product snippets reports in Search Console for errors by template, and Merchant Center diagnostics for price and availability mismatches.
- Monthly: crawl the product pages and compare the markup values with the feed, for every product if the crawl allows it.
- Outcomes: click-through rate and revenue for product pages from organic search, the number of disapproved items, and how AI answers describe your products for a fixed set of shopping questions. Our article on [Google AI Mode for e-commerce](https://marketingsignals.com/insights/google-ai-mode-for-ecommerce) explains how we track that last one.

For leadership, three numbers tell the story: the share of products with complete, valid markup, the number disapproved for data mismatches and the trend in organic revenue from product pages. They are operational measures, but each protects revenue directly.

## When it does not fit, and the mistakes we see

- Marking up content that is not on the page, such as ratings held by a review platform but not displayed. Google's guidelines say not to mark up content that is not visible to readers.
- Pulling third-party ratings into your own aggregateRating markup, which the review guidelines rule out.
- Product markup on category pages that list many products. Google's product rich results support pages that focus on a single product or variants of the same product.
- Two sources of markup on one page, such as a plugin and the theme, giving different prices or availability.
- Expecting markup to rank pages on its own. Google does not guarantee that structured data will show in results; it makes pages eligible, and the content still has to earn its place.
- Small catalogues with no feed: start with accurate pages and a Merchant Center account before working through every recommended property.

## Questions to ask your team or agency

- Do the price, availability, shipping and returns in our markup match the page and the feed for every product, and how do we know?
- Is our Product markup in the initial HTML, or added by JavaScript after the page loads?
- How are our variants marked up, and is there one canonical URL for each product group?
- Which of our policies are set in Merchant Center, Search Console or markup, and which one does Google use?
- Which facts that shoppers compare are missing from our product pages?

## Where to start

Our [e-commerce visibility playbook](https://marketingsignals.com/playbooks/the-e-commerce-visibility-playbook) shows where product data fits in the wider programme, with sector detail in the playbooks for [fashion e-commerce](https://marketingsignals.com/playbooks/the-visibility-playbook-for-fashion-e-commerce) and [marketplaces and aggregators](https://marketingsignals.com/playbooks/the-visibility-playbook-for-marketplaces-and-aggregators). If you want your product templates audited against Google's documentation and checked against how AI answers describe your range, our [AI search optimisation](https://marketingsignals.com/service/ai-search-optimisation) team can help.

## Frequently asked questions

### What is product schema?

Product schema is structured data, usually written in JSON-LD, that describes a product in a form search engines can read: its name, image, brand, identifiers such as GTIN, and a nested Offer with price, currency, availability, shipping and returns. Google uses it for product snippets and merchant listings in search results, and Merchant Center can use it to update price and availability automatically.

### Does product schema help with AI Overviews and AI Mode?

Google says no special schema is needed to appear in AI Overviews or AI Mode. Its guidance asks for structured data that matches the visible text and Merchant Center information kept up to date. Accurate product schema helps Google understand your products, but the facts an AI answer quotes still need to be written on the page, because that is what any assistant can read.

### Can I mark up reviews collected by a third-party review platform?

Google's review snippet guidelines say the review content you mark up must be readily available on the page, and that you should not aggregate reviews or ratings from other websites. If reviews are displayed on your product page and marked up to match, they can be eligible; ratings held elsewhere and not shown should not be added to your markup.

### Should shipping and returns go in schema, Merchant Center or Search Console?

Any of them can work, but Google applies an order of precedence: product-level Merchant Center feeds first, then Content API settings, then Merchant Center or Search Console settings, then product-level markup, then organisation-level markup. Pick one master source for each policy, keep the others consistent with it, and make sure the policy is also stated on the page.

### Do I need both structured data and a Merchant Center feed?

Google says providing both structured data on your pages and a Merchant Center feed maximises your eligibility for its shopping experiences, and some experiences combine data from both. For shipping and returns a product-level feed takes precedence over markup, while markup helps Google read your pages and lets Merchant Center make automatic price and availability updates. Both must match what shoppers see.

## Sources

- [Google Merchant Center Help: About structured data markup for Merchant Center](https://support.google.com/merchants/answer/6069143)
- [Google Search Central: Intro to Product structured data on Google (last updated 10 December 2025)](https://developers.google.com/search/docs/appearance/structured-data/product)
- [Google Search Central: AI features and your website (last updated 10 December 2025)](https://developers.google.com/search/docs/appearance/ai-features)
- [Google Search Central: General structured data guidelines (last updated 10 July 2026)](https://developers.google.com/search/docs/appearance/structured-data/sd-policies)
- [Google Search Central: Merchant listing (Product, Offer) structured data (last updated 8 September 2026)](https://developers.google.com/search/docs/appearance/structured-data/merchant-listing)
- [Google Search Central: Product variant structured data (ProductGroup, Product) (last updated 8 September 2026)](https://developers.google.com/search/docs/appearance/structured-data/product-variants)
- [Google Search Central: Merchant return policy structured data (last updated 8 September 2026)](https://developers.google.com/search/docs/appearance/structured-data/return-policy)
- [Schema.org: OfferShippingDetails](https://schema.org/OfferShippingDetails)
- [Google Search Central: Review snippet (Review, AggregateRating) structured data (last updated 8 September 2026)](https://developers.google.com/search/docs/appearance/structured-data/review-snippet)
- [Google Merchant Center Help: Product Ratings eligibility](https://support.google.com/merchants/answer/14549080)
- [Search Console Help: Shopping reports and tools](https://support.google.com/webmasters/answer/12660034)

Visibility Playbooks

## Related Playbooks

More Insights

## More articles we think you’ll like