Schema Markup For AI Search: Which Types Actually Get Picked Up

Schema markup for AI search won’t get your pages cited on its own. Google says there’s no special schema you need for AI Overviews or AI Mode, and an Ahrefs study of 1,885 pages that added JSON LD found no clear citation lift in AI Overviews, AI Mode or ChatGPT. What schema still does is help search engines understand who you are and what you sell, and Microsoft has said Bing uses it for its LLMs.

So the types worth adding are the ones that describe real entities and facts: Organization, Person with Article, Product with Offer and Review, LocalBusiness and BreadcrumbList. FAQPage and HowTo are the ones to stop chasing, because Google no longer shows either rich result. Below: what each platform has said, what the data shows, a verdict by type, and a setup order.

What Google, Bing And OpenAI Have Said About Schema Markup For AI Search

Most articles on this topic quote no one. Here’s what the platforms have actually published.

Google: helpful for Search, not required for AI. Google’s guide to optimizing for generative AI features says: “Structured data isn’t required for generative AI search, and there’s no special schema.org markup you need to add.” It adds that it’s still a good idea as part of your SEO, because it helps with rich results. Google’s AI features documentation makes the same point and lists one schema rule that matters: make sure your structured data matches the visible text on the page.

Microsoft Bing: schema helps its LLMs. At SMX Munich in March 2025, Bing’s Fabrice Canel confirmed that Microsoft uses schema markup to help its LLMs understand content. In October 2025, a Microsoft Bing product manager wrote that “schema is a type of code that helps search engines and AI systems understand your content” and named product, review, FAQ and event as examples.

OpenAI: no public statement on schema. OpenAI hasn’t documented whether ChatGPT search uses schema markup on web pages. For shopping, it runs a separate product feed specification that merchants submit directly, which tells you where it wants structured product data to come from.

The honest summary: one major engine says schema helps its AI understand pages, one says it isn’t needed for AI features, and one hasn’t said. None says schema earns citations.

What The Data Shows: 1,885 Pages Added Schema, Citations Barely Moved

The best test of schema markup for AI search so far is Ahrefs’ study of pages that added JSON LD, published in May 2026. It tracked 1,885 pages that added schema between August 2025 and March 2026 and compared them with about 4,000 similar pages that didn’t, 30 days before and after.

A first pass across 6 million URLs found AI cited pages were almost three times more likely to have JSON LD. That’s the number many vendors quote. The controlled test shows why it misleads: well run sites tend to have schema and also tend to get cited, but adding schema didn’t change citations.

Know the limits before you quote it. Every page in the study was already cited heavily by AI Overviews, so it says nothing about pages AI has never surfaced. It pooled all schema types together, used a 30 day window, and tested only JSON LD in the page HTML. Ahrefs still recommends schema for rich results, knowledge graphs and entity recognition.

Why Schema Added With JavaScript May Be Invisible To AI Crawlers

This is the most practical point about schema markup for AI search, and almost nobody covers it. Vercel’s analysis of AI crawler traffic found that none of the major AI crawlers it studied, including OpenAI’s and Anthropic’s, render JavaScript. They fetch JS files but don’t run them. Googlebot, and Gemini through Google’s infrastructure, do render it.

If your schema is added by Google Tag Manager or another script after the page loads, Google can see it, but ChatGPT’s and Claude’s crawlers likely can’t. To check, open your page, choose View Page Source (not Inspect), and search for “application/ld+json”. If it isn’t in the source, it’s being injected by script. Plugins like Rank Math print schema straight into the HTML, which avoids the problem.

Which Schema Types Actually Get Picked Up?

“Picked up” means two things here: whether Google still shows a search feature for the type (listed in its search gallery), and whether the type describes facts AI systems need to get your brand right.

The pattern for schema markup for AI search: types that describe who you are and what you sell keep their value; types that were mainly used to win extra SERP space have lost it.

The FAQ Schema Question: Rich Result Gone, Markup Still Valid

Google’s Search Central changelog records it plainly: on May 8, 2026 Google added a deprecation notice because the FAQ rich result “will no longer appear in Google Search starting May 7, 2026,” and on June 15, 2026 it removed the documentation. HowTo results were dropped back in 2023.

FAQPage is still a valid schema.org type, and leaving existing markup in place won’t hurt. But it’s no reason to add FAQs to pages that don’t need them. Write a FAQ section when readers genuinely have those questions; the visible answers are what AI systems can quote.

One rule matters more than the type you pick. Google’s structured data policies say: “Don’t mark up content that is not visible to readers of the page.” Putting facts in schema that aren’t on the page, such as a price or a claim, can cost you rich result eligibility, and it’s exactly the kind of mismatch that makes engines distrust a page.

Setting Up Schema Markup For AI Search, Step By Step

If you’re starting schema markup for AI search from scratch, do it in this order. Each step builds on the one before.

1. Organization on the homepage. Name, logo, URL, contact details and sameAs links to your official profiles. Give it a stable @id so other pages can reference it.
2. Person and Article on every post. Name the real author, link to their author page and LinkedIn, and connect the Article to your Organization as publisher.
3. Product or LocalBusiness where it fits. Prices, availability, address and hours must match what’s visible on the page.
4. BreadcrumbList sitewide. Most SEO plugins add it automatically.
5. Check it’s in the HTML source, not injected by script (see above).
6. Validate with Google’s Rich Results Test for supported features and the Schema Markup Validator for everything else.

Here’s what a basic Organization block looks like for a fictional company:

<script type=”application/ld+json”>

{

  “@context”: “https://schema.org”,

  “@type”: “Organization”,

  “@id”: “https://example.com/#organization”,

  “name”: “Northside Plumbing”,

  “url”: “https://example.com/”,

  “logo”: “https://example.com/logo.png”,

  “telephone”: “+1 602 555 0100”,

  “sameAs”: [

    “https://www.linkedin.com/company/example”,

    “https://www.facebook.com/example”

  ]

}

</script>

Google’s Organization documentation lists the full set of properties. In WordPress, Rank Math builds most of this from your settings: fill in your business details and social profiles there rather than pasting a second Organization block, which creates duplicates.

How To Tell If Your Schema Is Doing Anything

Be realistic about what schema markup for AI search can show you. Search Console shows rich result reports and errors for supported types, so you can confirm Google reads your markup. You can’t isolate a schema effect on AI citations from a single site, because citations move with content, links and platform changes at the same time.

What you can do is log the date you made schema changes and watch AI referral traffic in GA4 and your AI visibility tool for the following weeks. If nothing moves, that matches the study. Treat schema markup for AI search as hygiene that keeps your facts consistent, not as a growth lever, and put the bigger effort into pages that answer questions well. Our guides on how to rank on ChatGPT and getting cited in Google AI Overviews cover where that effort pays off.

Common Schema Mistakes That Hurt More Than They Help

• Markup that doesn’t match the page. Prices, ratings or FAQs in schema that readers can’t see break Google’s policies.
• Review markup for reviews you wrote yourself. Self serving or invented reviews in schema are a fast way to lose rich results.
• Two plugins outputting schema. Duplicate Organization or Article blocks with conflicting details confuse rather than clarify.
• Schema only in JavaScript. Fine for Google, likely invisible to most AI crawlers.
• sameAs links to dead or unofficial profiles. Point only to profiles you control and keep active.
• Treating schema as the AI strategy. Like llms.txt, it’s a small technical item, not a plan.

If you’d rather have the technical side and the content side handled together, that’s the AI search optimization work Rizqly does alongside SEO and paid media.

Frequently Asked Questions About Schema Markup For AI Search

Q1: Does Schema Markup Help With Google AI Overviews?

Not directly. Google says no special schema is needed for AI Overviews, and the Ahrefs study found no citation lift after pages added JSON LD. Schema still helps Google understand your pages and qualify for rich results.

Q2: Does ChatGPT Read Schema Markup?

OpenAI hasn’t said. Its crawlers read the initial HTML, so JSON LD printed in the source is available to them, but there’s no public confirmation of how it’s used. For shopping, OpenAI uses its own product feed.

Q3: Is FAQ Schema Still Worth Adding?

Not for the rich result, which Google removed on May 7, 2026. Keep existing markup if the FAQs are visible on the page, but don’t add FAQs just to have schema.

Q4: Which Format Should I Use For Schema Markup?

JSON LD. Google recommends it as the easiest to implement and maintain, and it keeps markup separate from your page design. It’s also the format most guides on schema markup for AI search assume.

For the full picture of how AI search engines choose sources, start with our guide to generative engine optimization.