flareOps
Guide

Product schema for AI shopping: which fields matter and how to check them

Last checked: · Updated · By flareOps (how we write)

In short: Product structured data (JSON-LD) tells machines your product's name, image, price, currency, stock, brand and GTIN. Google requires name, image and offers with price and currency, and recommends availability, brand, GTIN, shipping and returns. Put it in the first HTML, not JavaScript, and match the visible page. It helps AI read you; it does not guarantee a recommendation.

Product structured data is a small block of code on each product page that says, in a format machines understand, "this is a product, it costs this much, it is in stock, it is made by this brand." Google reads it. Shopping feeds copy from the same facts. AI assistants that fetch your page can read it too. This guide covers what to include and how to check it on WooCommerce, Wix and custom sites. Facts were checked on 4 Nov 2026.

First, an honest expectation

Google's 2026 guide on generative AI search says: "Structured data isn't required for generative AI search, and there's no special schema.org markup you need to add." It also asks that your structured data match the visible text on the page.

So Product schema is hygiene, not a hack. It makes your facts easy to read and hard to misread. It does not make AI recommend you on its own.

The fields that matter

Google's merchant listing documentation splits the fields like this:

FieldGoogle saysWhy it matters for AI shopping
nameRequiredThe product's identity
imageRequiredShown in shopping results
offers.priceRequiredPrice is a top factor; OpenAI lists price among how it ranks merchants
offers.priceCurrencyRequiredAvoids wrong currency in answers
offers.availabilityRecommendedOpenAI lists availability among how it ranks merchants
brand.nameRecommendedMatches your product to a known item
gtin (or gtin8/12/13/14)RecommendedThe strongest way to match the exact product
mpn, skuRecommendedBackup identifiers
descriptionRecommendedFacts for questions about use and size
offers.shippingDetailsRecommendedAnswers "does it ship to…, how fast"
offers.hasMerchantReturnPolicyRecommendedAnswers "can I return it"
color, material, sizeRecommendedMatches the small searches AI makes behind one question
aggregateRating, reviewRecommendedOnly with real reviews shown on the page

If you use shippingDetails, Google needs shippingDestination, deliveryTime and shippingRate. If you use hasMerchantReturnPolicy, it needs applicableCountry and returnPolicyCategory, plus merchantReturnDays for a fixed return window.

A clean example

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Product",
  "name": "Ceramic Planter with Drainage Hole, 15 cm",
  "image": "https://example.com/img/planter-15.jpg",
  "description": "Matte glazed ceramic planter, 15 cm wide, drainage hole and saucer included.",
  "sku": "CP-15",
  "brand": {"@type": "Brand", "name": "Example Pottery"},
  "offers": {
    "@type": "Offer",
    "url": "https://example.com/products/ceramic-planter-15",
    "price": "38.00",
    "priceCurrency": "USD",
    "availability": "https://schema.org/InStock",
    "itemCondition": "https://schema.org/NewCondition"
  }
}
</script>

This made-up example has no gtin because it is the store's own brand with no barcode. That is correct. Never add a GTIN you do not have.

The JavaScript price trap

This is a common problem on small stores, and it is invisible in a normal browser.

Some themes, page builders and custom sites add the price, the stock line or the whole JSON-LD block after the page loads, using JavaScript. Your browser runs the script, so the page looks fine. But:

  • many AI fetchers read only the HTML the server sends;
  • Google's merchant listing docs warn that "dynamically-generated markup can make Shopping crawls less frequent and less reliable", and recommend putting Product structured data "in the initial HTML for best results."

How to check: open the product page, choose "View page source" (not "Inspect"), and search for the price. If it is not there, AI fetchers may not see it either.

Other traps

  1. Two Product blocks. WooCommerce, your theme and an SEO plugin (Yoast, Rank Math) may each print one. Different prices in different blocks confuse every reader. Keep one.
  2. Price in the markup differs from the page. Sale prices and currency-switcher plugins cause this. Google may disapprove the product in Merchant Center.
  3. Stock says InStock when the page says sold out. Usually a cache. Clear it after stock changes.
  4. Variations. For a product with sizes or colours, either one Product with several offers or a ProductGroup with variants. Make sure the price shown matches the selected or lowest variant consistently.
  5. Fake ratings. Stars without real, visible reviews break Google's rules.

Platform notes

  • WooCommerce: WooCommerce outputs Product JSON-LD by default. SEO plugins often extend it. WooCommerce 9.2 and later has a built-in GTIN field, but your markup only includes the GTIN if a plugin or theme maps it. See the WooCommerce GTIN guide.
  • Wix: Wix adds Product JSON-LD automatically. In our own tests on Wix dev stores, its default product markup did not include GTIN or MPN. Wix lets you edit SEO structured data per page.
  • Custom sites: generate the JSON-LD on the server from the same data that draws the page, so they cannot drift apart. In Next.js, render it in the page component, not in a client-only effect. See AI shopping for a custom site.

How to test

  1. View page source and search for the price and for "@type": "Product".
  2. Run Google's Rich Results Test on the URL.
  3. Paste the URL into the Schema.org validator.
  4. Run the free AI readiness checker: it fetches the page as a plain client, without JavaScript, the way many AI fetchers do.

How flareOps helps

flareOps' daily store check reads each product page the way a plain fetcher does and flags: no Product JSON-LD, JSON-LD that fails to parse, more than one Product block, missing or mismatched price, currency or availability, brand or GTIN known but missing from the markup, price not in the server-sent text, stock not stated in the page, and pages that need JavaScript to show anything. GTINs are checked with their check digit. With the WooCommerce plugin or the @flareops.ai/next package on a custom site, approved fixes can be written into the page markup, with undo. On Wix, approved changes are written into the Wix product; GTIN in Wix's markup needs an extra Wix permission. For short names, short descriptions and missing brands, flareOps gives you a ready prompt for your AI assistant, using only facts you confirm. flareOps never invents a GTIN, price or stock status. The store check is free; Pro is $29 per month per store.

Questions people ask

Is Product schema required to appear in Google AI Mode?

No. Google says structured data is not required for generative AI search and there is no special markup to add. It still helps Google read prices and stock correctly, and it is required for some merchant listing features.

Does ChatGPT read JSON-LD?

OpenAI says ChatGPT uses structured product metadata such as price and description from first-party and third-party providers. OpenAI does not publish exactly how it parses page markup, so treat JSON-LD as helpful, readable data rather than a guaranteed input.

What is the JavaScript price trap?

Some themes and site builders add the price, stock or the whole JSON-LD only after JavaScript runs. Many AI fetchers do not run JavaScript, and Google warns that dynamic markup can make Shopping crawls less frequent and less reliable. Put the price in the server-sent HTML.

My page has two Product blocks. Is that a problem?

It often is. A theme, WooCommerce and an SEO plugin can each output one, sometimes with different prices. Keep one complete Product block per product page.

Should I add reviews and ratings to my schema?

Only real reviews that are shown on the page. Never add made-up ratings to get stars. Google treats markup that does not match the page as a policy problem.

What if my product has no GTIN?

Leave gtin out. Use brand and mpn if you have a real manufacturer part number. For your own handmade brand, no GTIN is normal.

Which tool should I use to test my markup?

Google's Rich Results Test shows what Google can read, and the Schema.org validator checks the syntax. Also view the page source (not the inspector) to confirm the price is in the first HTML.

Sources

  1. Google Search Central: Merchant listing structured data
  2. Google Search Central: AI features and your website
  3. Google Search Central: Optimizing for generative AI features on Google Search
  4. OpenAI Help: Shopping in ChatGPT
  5. Google: Rich Results Test
  6. Schema.org: Product
  7. Schema.org validator
  8. Google Search Central: FAQPage structured data (rich result deprecated 2026)

Sources checked on 4 November 2026.