Structured data for AI search: what schema markup helps, and what it cannot do

Schema markup helps AI search in one modest, useful way: it states facts about your organisation and your pages in a form machines can read without guessing. It does not earn AI citations on its own, and Google says there is no special schema you need for its AI features.

Schema markup, also called structured data, is a block of code in a page's HTML that labels what the page contains: this is an organisation, this is its logo, this article was published on this date by this author. The labels come from the shared schema.org vocabulary, and Google recommends writing them as JSON-LD "if your site's setup allows it". This post covers which types are worth adding for AI search, what the official guidance says, and where the claims made for schema go too far.

In short

• Google says "structured data isn't required for generative AI search, and there's no special schema.org markup you need to add", but recommends keeping it for rich results in ordinary search.
• OpenAI, Anthropic and Perplexity say nothing about structured data in their crawler documentation. There is no published evidence that schema makes an assistant cite you.
• The schema worth having: Organization (with sameAs links) on your home page, and Article or BlogPosting on articles.
• Google stopped showing FAQ rich results on 7 May 2026. FAQPage markup is still valid, but it no longer earns a special display in Google.
• Every value in your markup must match what the page visibly says, and the markup must be in the HTML your server sends.

Does schema markup help in AI search?

Google's answer is "not specially". Its guide to generative AI features in Search lists "overfocusing on structured data" among the things you can ignore. Structured data "isn't required for generative AI search", there is "no special schema.org markup" to add, and it remains "a good idea" as part of SEO "as it helps with being eligible for rich results". Google's AI features page adds one practical rule: make sure your structured data matches the visible text on the page.

The other assistants say nothing. OpenAI's, Anthropic's and Perplexity's crawler pages, checked on 11 October 2026, do not mention structured data at all. They may read it; they may not. Nobody has published how, or whether, it affects which brands their answers name.

What schema does reliably do is remove ambiguity. Google says Organization markup on your home page "can help Google better understand your organization's administrative details and disambiguate your organization in search results". If your brand shares a name with another business, that is worth having. It is the same problem AI assistants have when they mix two similarly named companies, covered in how to stop AI confusing your brand.

So treat schema as accurate labelling, not a ranking lever. The words on the page still do the work: see how to write content AI assistants quote.

Which schema markup matters for AI search?

Type or property	Where it goes	What Google's documentation says it does	Status, October 2026
Organization	Home page	Helps Google understand your administrative details, disambiguate your organisation, and choose the logo shown in results	Current. No required properties
sameAs	Inside Organization or an author	Links to pages on other sites about the same organisation, such as a social media or review profile	Current. Multiple URLs allowed
Article or BlogPosting	Each article or blog post	Helps Google show better title text, images and dates for the article	Current. No required properties
FAQPage	Pages with a visible FAQ	Used to produce the FAQ rich result	Rich result no longer shown in Google Search since 7 May 2026

For Article and BlogPosting, Google recommends marking up the author, with a URL that identifies them, such as a bio page or social profile. It says it can "understand both sameAs and url when disambiguating authors".

How does sameAs help AI tell your brand apart?

schema.org defines sameAs as the "URL of a reference Web page that unambiguously indicates the item's identity", for example a Wikipedia page, a Wikidata entry or an official website. Google's Organization documentation gives social media and review site profiles as examples.

For a South African business, that usually means the official LinkedIn, Facebook, Instagram or YouTube pages you actually run, plus any directory or review profiles that are yours. For example, a fictional Johannesburg insurer's home page might carry:

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "Kestrel Insurance",
  "url": "https://www.example.com",
  "logo": "https://www.example.com/images/logo.png",
  "description": "Kestrel Insurance sells car, home and business insurance in South Africa.",
  "email": "hello@example.com",
  "address": {
    "@type": "PostalAddress",
    "addressLocality": "Johannesburg",
    "addressCountry": "ZA"
  },
  "sameAs": [
    "https://www.linkedin.com/company/YOUR-PAGE",
    "https://www.facebook.com/YOUR-PAGE"
  ]
}

Only list profiles you control and that describe the same business. A sameAs link to the wrong profile, or to a company with a similar name, adds confusion instead of removing it.

Is FAQPage schema still worth adding?

Not for Google's display any more. Google added a deprecation notice on 8 May 2026 saying the FAQ rich result "will no longer appear in Google Search starting May 7, 2026", and removed the documentation in June 2026. The FAQPage type itself still exists in the schema.org vocabulary.

This blog still publishes FAQPage markup alongside each post's visible FAQ. It costs nothing and describes the page accurately. But the value sits in the visible questions and answers: short, self-contained answers that a reader, or an AI system reading the page, can lift whole. Write the FAQ for people first. The markup is a label on it, not the reason it works.

What rules does Google set for structured data?

Google's general structured data guidelines apply to every type. The ones that matter most:

1. Mark up only what readers can see. Google says not to mark up content that is not visible on the page, and lists content "hidden from the user" among the reasons markup may not be shown.
2. Keep it accurate. No misleading content, fake reviews or impersonation. A structured data issue can lead to a manual action, which removes the page's eligibility for rich results but, Google says, "doesn't affect how the page ranks".
3. Expect no guarantee. Google "does not guarantee that your structured data will show up in search results", even when it is correct.
4. Test it. Google's Rich Results Test and the URL Inspection tool in Search Console catch most technical errors.
5. Put it in the server's HTML. Google can process markup added by JavaScript, but a crawler that does not run JavaScript will never see it. Check with the curl test in is your website readable by ChatGPT?

Schema markup is often sold as a shortcut to AI visibility, along with robots.txt and llms.txt. robots.txt controls access, Google says its Search ignores llms.txt, and schema labels facts. None of them replaces a page that answers the question well.