Schema markup for AI search: what matters in 2026

In 2026, no schema markup will get a page into AI search by itself. Google says AI Overviews and AI Mode need no special structured data, and a controlled Ahrefs study of 1,885 pages found that adding JSON-LD did not lift AI citations. Schema still pays where Google displays it: prices, stock, review stars, events, videos, breadcrumbs and your business details.

Schema markup, or structured data, is code that labels what a page contains, such as a product and its price, using the shared schema.org vocabulary. Most sites add it as JSON-LD, a block of code visitors never see. A rich result is a search listing Google decorates with that data, like star ratings or a price.

Does schema markup help with AI search in 2026?

Schema markup has no proven direct effect on AI search in 2026. Google's guide to AI features in Search says there's "no special schema.org structured data that you need to add" to appear in AI Overviews, the AI summary at the top of some results, or AI Mode, its chat-style search.

A page only needs to be indexed and eligible to show with a snippet, which is plain SEO (see what still works in SEO in 2026). Google's AI optimization guide, added on May 15, 2026, lists "overfocusing on structured data" among things to ignore, yet still recommends it for rich results. The AI features page adds that markup should match the visible text. Google's case for schema is rich results, not AI answers.

Microsoft goes further. Fabrice Canel of Bing said at SMX Munich in March 2025 that schema markup helps Microsoft's language models understand content, Search Engine Land reported, and Bing's October 2025 guide to AI answers says the same. That covers Bing and Copilot, nothing else.

Google's John Mueller, in a Reddit reply reported on January 2, 2026, offered a personal view, not official guidance: "yes, no, and it depends." His clearest yes was shopping, where he said prices, shipping and availability are basically impossible to read accurately from page text.

Which rich results did Google remove in 2025 and 2026?

Google has removed eleven structured data features from its results since 2023, and FAQ rich results were the latest, gone for every site on May 7, 2026. Leftover markup simply stops changing how a result looks. Dates are from Google's documentation changelog and blog.

Book actions was on the June 2025 list too, but Google kept it. For FAQ, Google's notice said its Search Console report and Rich Results Test support would go in June 2026, and API support in August 2026.

Is FAQ schema still worth adding in 2026?

FAQ schema earns nothing visible in Google in 2026: the dropdown questions under results stopped on May 7 for every site, government and health sites included. Google said in 2023 that unused markup causes no problems, so there's no need to strip it out, but new FAQPage markup added for Google is wasted effort.

The questions themselves still matter. Bing's guide notes that assistants can often lift question and answer pairs word for word, a benefit that comes from the visible text, not the markup. We found no published test that isolates FAQPage markup and shows a gain in AI citations.

Do ChatGPT, Perplexity and Gemini read schema markup?

Whether ChatGPT, Perplexity and Gemini read schema markup is unknown outside those companies, and only Microsoft has said schema helps its AI. OpenAI says ChatGPT's shopping results draw on structured product data from merchants and data providers, but we found no statement from OpenAI, Perplexity, Anthropic or Google's Gemini app saying they parse the JSON-LD on ordinary pages. Small tests suggest assistants read it as plain text, if at all.

What is known

• When ChatGPT picks products, it considers "structured metadata from first-party and third-party providers", such as price and description, according to OpenAI's shopping help page. Shopify stores are included through Shopify Catalog; other merchants can apply to send a product feed.
• Google's 2026 guide says Merchant Center feeds and Business Profiles help products and local businesses appear in AI responses. Those are feeds and listings, not page markup.

What the tests found

In October 2025, SEO software company searchVIU placed a fictional product's prices in eight different ways, from visible text to JSON-LD only. None of the five systems it asked, ChatGPT, Claude, Gemini, Perplexity and Google AI Mode, returned the price that appeared only in JSON-LD.

In February 2026, SEO consultant Mark Williams-Cook put a made-up company's address only inside deliberately invalid JSON-LD. ChatGPT and Perplexity returned it anyway. His reading: they treat the code as more text, not as structured data. He still recommends schema.

The biggest controlled study we found, published by Ahrefs on May 11, 2026, tracked 1,885 pages that added JSON-LD between August 2025 and March 2026 against 4,000 control pages that didn't, for 30 days either side.

• Google AI Overviews: citations fell 4.6% against the controls, a small but statistically significant gap Ahrefs couldn't pin on schema.
• Google AI Mode: up 2.4%, indistinguishable from zero.
• ChatGPT: up 2.2%, indistinguishable from zero.

What is still a guess

• Whether schema helps a page AI hasn't found yet. Every page in the Ahrefs study was already cited heavily.
• Whether some types help more than others; Ahrefs pooled them.
• Whether schema in training data counts. Williams-Cook argues training pipelines strip it out.

The fair summary for October 2026: for AI search, schema is housekeeping. What it reliably earns is Google's rich results, and Microsoft says it helps Bing's AI. For what gets pages cited, see how to rank in ChatGPT and AI search in 2026 and how to appear in Google AI Overviews.

Which schema types matter in 2026?

For most business sites in 2026, seven schema types still do real work: Organization, LocalBusiness, Product with Offer, Review, Article, BreadcrumbList and Event, plus VideoObject for sites with video. The table pairs what Google shows for each, per its search gallery on October 25, 2026, with an honest read of its AI value.

Google is still adding to these types: product category and sale dates in July 2026, and a video creator property in September.

Where to start: Organization on the home page, then LocalBusiness or Product, then BreadcrumbList. After that, an hour on your Business Profile or Merchant Center feed beats another schema type.

How do you add schema markup the right way?

Schema markup goes on the page it describes, as JSON-LD, the format Google recommends as easiest to maintain. On WordPress, Shopify or Wix, a plugin or the platform's SEO settings may already write it, so check what your site outputs before adding more.

1. Mark up only what visitors can see. If the JSON-LD names a price, the page must show that price.
2. Put it in the HTML your server sends. Google reads JSON-LD added by JavaScript, but Vercel's December 2024 data found crawlers from OpenAI, Anthropic and Perplexity didn't run JavaScript. Our list of AI crawlers covers who visits.
3. Fill fewer properties well. Google prefers complete, accurate data to every property filled badly.
4. Add sameAs to Organization with links to your official profiles.
5. Update the markup when prices, stock or hours change.
6. Don't mark up your own testimonials. Google gives no stars to LocalBusiness or Organization pages when the business controls the reviews about itself.

Bad markup can cost a page its rich results through a manual action, but not its rankings, Google's policy says. Correct markup doesn't guarantee a rich result either.

How to test schema markup

Test schema markup in three places: Google's Rich Results Test before publishing, the Schema Markup Validator for any schema.org markup, and Search Console after release. The first two take a URL or code; Search Console tracks errors across the site.

1. Run the page through the Rich Results Test. It lists the rich result types it found, with errors and suggestions.
2. Check the code in the Schema Markup Validator, which also reads types Google doesn't use.
3. Once live, use URL Inspection in Search Console to see what Google fetched.
4. Watch the rich result reports under Enhancements, or Shopping for products. They show a sample of items, not every page, and Google drops a report when it retires the feature, as it planned for FAQ in June 2026.

How QuQi helps

Structured data is part of what QuQi's crew of AI agents works on. The SEO Auditor audits your site, puts what costs the most traffic first and tells the Page Optimizer what to fix. The Page Optimizer writes fixes for titles, descriptions and structured data on the pages the audit flags, the most valuable pages first.

AI Visibility checks every week whether Google's AI Overview, ChatGPT, Perplexity and Gemini name your business for your buyers' questions, and records who is named instead. Every batch runs in a chat you can open, and the approval switch for SEO fixes decides whether they go out on their own or wait for your OK. Work is paid in credits (see pricing); QuQi doesn't promise rankings or citations.

Questions people ask