Schema Markup for AI Search: What It Does and What It Doesn’t

Schema markup for AI search is structured data (usually JSON-LD using the schema.org vocabulary) that labels the facts on your page so machines can read them without guessing. Here’s the honest part. Google says structured data isn’t required to appear in AI Overviews or AI Mode, and there’s no special AI schema to add. It still helps Google understand your pages and makes them eligible for rich results, so it’s worth doing well. It just isn’t a shortcut into AI answers.

I’ll say that plainly because a lot of agencies sell schema as the key to getting cited by ChatGPT. I haven’t found a primary source from OpenAI, Anthropic or Perplexity that says so. So in this guide I’ll stick to what’s documented: what structured data really does, which types matter for a business in September 2026, which ones to stop adding, a JSON-LD example, and how I test markup before it goes live.

What Does Structured Data Actually Do in Google Search?

Two jobs, per Google’s own introduction to structured data. The first is understanding. Google says it uses structured data it finds on the web to understand the content of a page, and to gather information about the world, such as the people, books or companies in the markup.

The second is rich results. Markup can make a page eligible for enhanced search listings like product details, review stars, event dates or recipe cards. Google’s page includes its own case studies; the Rotten Tomatoes one reports a 25% higher click-through rate on pages with structured data compared with pages without it.

Notice the word “eligible”. Markup never guarantees a rich result, and Google doesn’t treat it as a ranking boost. In my experience that’s the single most misunderstood point in the whole topic. Schema describes what’s on the page. It doesn’t make the page better.

What Does Google Say About Schema Markup for AI Search?

Google has written about this directly, twice. Its AI features page says you don’t need new machine readable files, AI text files or markup to appear in AI Overviews and AI Mode, and adds that “there’s also no special schema.org structured data that you need to add.” The same page lists “making sure your structured data matches the visible text on the page” among the SEO basics that still apply.

Then in May 2026 Google published its guide to optimizing for generative AI features. It has a mythbusting section, and one heading is literally “Overfocusing on structured data”. The wording: structured data isn’t required for generative AI search, and there’s no special schema.org markup you need to add. Google then says it’s still a good idea to keep using it as part of your SEO strategy, because it helps with rich result eligibility.

My reading is simple. Google’s AI features pull from the same index as classic search, and the guide says a page must be indexed and eligible for a snippet to show up in them. So schema helps the way it always has: clearer understanding, cleaner listings. Nothing more mystical than that.

Does Schema Markup Help You Get Cited in ChatGPT or Perplexity?

Nobody who runs those products has said so in their documentation. I checked OpenAI’s crawler page and Perplexity’s bot guide in September 2026. Both talk about letting their search crawlers (OAI-SearchBot and PerplexityBot) reach your pages through robots.txt. Neither mentions schema as a citation factor, and I list what OpenAI does document in my guide on how to get cited by ChatGPT.

The closest thing to an official statement came from Microsoft. At SMX Munich in March 2025, Bing’s Fabrice Canel said schema markup helps Microsoft’s LLMs understand content, as reported by Search Engine Land. That’s a conference remark, not documentation, and it’s about understanding, not a promise of citations in Copilot.

So what do I tell clients? Add accurate markup because it helps search engines read your facts, and because clean facts can only help any system that parses your page. Don’t pay anyone who claims a specific schema type will get you quoted by ChatGPT. That claim has no primary source behind it. For the wider question of how AI visibility differs from classic SEO, see my piece on GEO vs SEO.

Which Schema Types Matter for Your Business?

This is the decision that actually matters. Three variables change the answer:

1. What you sell. Physical products, local services or information.
2. Where you sell it. A storefront or service area, or online only.
3. What’s visible on the page. Markup can only describe what a visitor can see, so thin pages get thin markup.

If you’re a local business, then add LocalBusiness on your home or location page, using the most specific subtype schema.org offers (Plumber, Dentist, Restaurant and so on). Google’s docs list name and address as required, with hours, phone, geo coordinates and URL recommended. Pair it with a complete Google Business Profile, because Google’s AI features page lists an up-to-date Business Profile among the basics too. Consistent details across the web matter as much as the markup; my NAP consistency guide covers that side.

If you run an online store, then add Product markup to product pages. Google separates two uses: merchant listings for pages where people can buy, and product snippets for pages like editorial reviews where they can’t. Google also accepts product data through Merchant Center feeds and says some experiences combine both sources, so for stores I’d set up both, then use my product schema markup guide to see which class each page falls into.

If you publish articles, then add Article or BlogPosting markup. Google says there are no required properties for Article. The ones I never skip are headline, dates, image and author, with author.url pointing to a real bio page and only the person’s name in author.name.

Whatever you are, add Organization markup once, on the home page or about page. I cover that type in depth in my entity SEO guide, so I won’t repeat it here.

Which Schema Types Should You Stop Adding?

This is where old checklists hurt people. Google has retired several rich results, and markup for a retired feature earns nothing in Search.

My rule: I don’t add new HowTo or FAQ markup for rich results, full stop. Existing FAQ markup won’t hurt you, so there’s no emergency. Just don’t spend a developer’s afternoon adding more of it because a 2023 blog post said so.

A JSON-LD Example You Can Adapt

Here’s LocalBusiness markup for an invented plumbing company, using the Plumber subtype. Replace every value with facts that appear on your own page. If your hours or phone number on the page differ from the markup, fix the page or the markup until they match.

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Plumber",
  "@id": "https://www.example.com/#business",
  "name": "Example Plumbing Co",
  "url": "https://www.example.com/",
  "telephone": "+1-512-555-0142",
  "image": "https://www.example.com/images/storefront.jpg",
  "priceRange": "$$",
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "123 Example Street",
    "addressLocality": "Austin",
    "addressRegion": "TX",
    "postalCode": "78701",
    "addressCountry": "US"
  },
  "geo": {
    "@type": "GeoCoordinates",
    "latitude": 30.26715,
    "longitude": -97.74306
  },
  "openingHoursSpecification": [
    {
      "@type": "OpeningHoursSpecification",
      "dayOfWeek": ["Monday", "Tuesday", "Wednesday", "Thursday", "Friday"],
      "opens": "08:00",
      "closes": "18:00"
    }
  ]
}
</script>

A few choices here are deliberate. The geo coordinates use five decimal places, which Google’s LocalBusiness docs ask for. There’s no aggregateRating, because Google lists review properties for sites that collect reviews about other businesses, so marking up your own testimonials won’t earn stars. And there’s nothing AI-specific in it. There doesn’t need to be.

How Do You Test Schema Markup Before and After Publishing?

I use three checks, in this order. Each catches something the others miss.

1. Rich Results Test (search.google.com/test/rich-results). Paste a URL or code. It tells you which Google rich results the page is eligible for and flags errors. It only reports on rich result types Google supports, so it won’t comment on every type in your markup.
2. Schema Markup Validator (validator.schema.org). This checks your markup against the full schema.org vocabulary, whatever Google supports. It’s the right tool for types Google doesn’t display, like Organization details or a niche LocalBusiness subtype.
3. Search Console. After publishing, use URL Inspection to see what Google read from the live page, and watch the rich result reports for errors across the site.

The most common problem I find isn’t a syntax error. It’s duplication. An SEO plugin outputs one Organization block, the theme outputs another, and a review widget adds a third with different names. Pick one source, switch the others off, and retest.

The Decision Matrix

If you only read one part of this guide to schema markup for AI search, read this.

When Should You Hand This to a Developer?

If your site runs on a mainstream CMS with a good SEO plugin, you can usually handle Article and Organization yourself. Bring in a developer, or a technical SEO specialist, when markup has to be generated from a product database, when you run many locations, or when your pages build content with JavaScript and you’re not sure what Google sees.

If you’d like a second pair of eyes on the whole setup, our AI search optimization service starts with a free audit, and schema is one of the things I check. And if you’re weighing the other “AI file” people talk about, my companion post on llms.txt explains what it does and who reads it.

Frequently Asked Questions

Is There a Special Schema for AI Overviews?

No. Google’s documentation says there’s no special schema.org structured data you need to add to appear in AI Overviews or AI Mode. Pages need to be indexed and eligible for a snippet. Regular structured data still helps Google understand the page and qualify for rich results.

Should I Remove My Existing FAQ Schema?

You don’t have to. Google stopped showing FAQ rich results on May 7, 2026, so the markup no longer earns a visual feature in Search, but Google hasn’t said it causes harm. I wouldn’t add new FAQ markup for rich results, and I’d remove it only when you’re already cleaning up a template.

Does Schema Markup Improve Rankings?

Google doesn’t describe structured data as a ranking factor. It describes it as a way to help Google understand a page and to make the page eligible for rich results. Better listings can earn more clicks, which is the real benefit.

What Is the Best Format for Schema Markup?

JSON-LD. Google supports JSON-LD, Microdata and RDFa, and recommends JSON-LD because it’s the easiest to add and maintain at scale. It sits in its own script block, so you can update it without touching the visible HTML.

Last updated: September 2026 by Mizanur Rahman