Technical AEO

    Schema Markup for AEO: What Actually Helps AI Understand Your Pages

    Which structured data types matter for AI search, what schema can and can't do for ChatGPT, Gemini and AI Overviews, and ready-to-use JSON-LD examples.

    Saarthak Srivastav

    Saarthak Srivastav

    · 5 min read

    Key takeaways

    • Schema markup helps search engines understand entities and page types. AI assistants that ground answers in those search engines benefit indirectly.
    • There's no evidence that schema alone gets you cited. Treat it as a foundation that removes ambiguity, not a ranking hack.
    • The highest-value types for AEO are Organization, Product with Offers, Article, and LocalBusiness, plus BreadcrumbList for site structure.
    • Structured data must match visible content. Marking up things users can't see risks manual actions and erodes trust.
    • FAQPage markup no longer earns rich results for most sites, but question-and-answer content on the page still helps AI retrieval.

    Schema markup helps AI search by making your pages unambiguous to the search systems that AI assistants rely on. It tells machines what a page is (a product, an article, a local business), who is behind it, and how it relates to other entities. It won't get you cited on its own; nothing on the page is a substitute for a clear, useful answer. But it removes guesswork, and in AEO ambiguity is expensive.

    What it can do:

    • Clarify entities. Organization markup with sameAs links tells search engines that your site, your LinkedIn page and your Crunchbase profile are the same company. That consistency feeds the knowledge graphs that AI features draw on.
    • Expose facts in a machine-readable way. Prices, availability, ratings, opening hours, authors and dates in structured form are less likely to be misread.
    • Describe site structure. BreadcrumbList shows how pages relate.
    • Make you eligible for rich results in traditional search, which can improve the visibility that AI retrieval builds on.

    Google states that structured data isn't required to appear in AI Overviews or AI Mode, but that it's useful for helping its systems understand content. Microsoft's Bing team has also spoken publicly about structured data helping its systems, which ground Copilot answers, understand pages.

    What it can't do:

    • It doesn't force a citation. AI answers are written from page content. If the visible text doesn't answer the question, markup won't save it.
    • It doesn't work as a direct LLM input in most cases. Many AI crawlers read the raw HTML, and schema is just one part of it. How each assistant uses JSON-LD isn't documented, so don't build your strategy on it.
    • It can't make up for poor content or blocked crawlers. Fix crawler access and content structure first.

    The schema types that matter most

    Organization (every site)

    This is the most important markup for AEO. Put it on your homepage or about page.

    {
      "@context": "https://schema.org",
      "@type": "Organization",
      "name": "Example Analytics",
      "url": "https://www.example.com",
      "logo": "https://www.example.com/logo.png",
      "description": "Example Analytics is a product analytics platform for mobile app teams.",
      "foundingDate": "2019",
      "sameAs": [
        "https://www.linkedin.com/company/example-analytics",
        "https://x.com/exampleanalytics",
        "https://www.crunchbase.com/organization/example-analytics",
        "https://github.com/example-analytics"
      ],
      "contactPoint": {
        "@type": "ContactPoint",
        "contactType": "sales",
        "email": "[email protected]"
      }
    }
    

    Tips:

    • Use the exact same description here as in your social bios and directory listings. Consistency is the point.
    • sameAs should list only official profiles you control, plus Wikipedia or Wikidata entries if they exist.
    • Use a more specific subtype where one fits: LocalBusiness, SoftwareApplication publisher, NewsMediaOrganization, EducationalOrganization and so on.

    Product and Offer (ecommerce and software)

    For product pages, mark up name, description, brand, price, currency, availability and, if genuine, aggregate ratings.

    {
      "@context": "https://schema.org",
      "@type": "Product",
      "name": "Trailhead 2 Hiking Boot",
      "description": "Waterproof leather hiking boot with a wide toe box, for day hikes and light backpacking.",
      "brand": { "@type": "Brand", "name": "Example Outdoor" },
      "sku": "TH2-BRN",
      "offers": {
        "@type": "Offer",
        "price": "149.00",
        "priceCurrency": "USD",
        "availability": "https://schema.org/InStock",
        "url": "https://www.example.com/boots/trailhead-2"
      }
    }
    

    AI shopping features rely heavily on accurate product data, both from markup and from merchant feeds. Keep prices and stock in sync with what's on the page.

    For software, SoftwareApplication with applicationCategory, operatingSystem and offers describes your product clearly.

    Article and BlogPosting (content)

    Mark up headline, author, publisher, datePublished and dateModified. Dates matter: AI systems increasingly weigh freshness for time-sensitive questions, and a visible, accurate modified date supports that.

    LocalBusiness (local companies)

    Name, address, phone, opening hours, geo-coordinates, area served and price range. Keep it identical to your Google Business Profile and major directories. Local AI answers ("best plumber near me open now") depend on these facts being consistent.

    Shows where a page sits in your site. Cheap to add and helps systems understand topical structure.

    FAQPage and HowTo

    In 2023 Google limited FAQ rich results to well-known, authoritative government and health sites, and stopped showing HowTo rich results. For most sites these types no longer produce a visual search feature.

    That doesn't make question-and-answer content less useful. Self-contained Q&A pairs are among the easiest passages for AI retrieval to match. Keep writing them where they answer real questions. Whether you also add FAQPage markup is low stakes either way.

    Implementation rules

    1. Use JSON-LD in a <script type="application/ld+json"> tag.
    2. Server-render it. Markup injected by JavaScript after load may be missed by crawlers that don't execute scripts, which includes most AI crawlers.
    3. Match visible content. Every marked-up fact should be visible on the page. Hidden or misleading markup violates Google's guidelines and can lead to manual actions.
    4. Be specific. Use the most specific type that applies, and fill in recommended properties, not just required ones.
    5. Use stable IDs. Give your Organization an @id (for example https://www.example.com/#organization) and reference it from other markup, so all your pages point to the same entity.
    6. Validate. Use Google's Rich Results Test and the Schema Markup Validator after every template change.
    7. Keep it current. Stale prices and discontinued products in markup cause the same problems as stale text.

    A practical rollout order

    Priority Markup Where
    1 Organization with sameAs Homepage or about page
    2 BreadcrumbList All templated pages
    3 Product / SoftwareApplication + Offer Product and pricing pages
    4 Article / BlogPosting with dates and author Blog and resources
    5 LocalBusiness Location pages
    6 Review / AggregateRating (only if genuine and on-page) Product pages

    How to check whether it's working

    Schema's effect on AI answers is indirect, so measure the outcomes it supports:

    • Rich result reports in Search Console for eligible types
    • Entity accuracy in AI answers: do assistants get your category, founding date, pricing model and location right?
    • Product fact accuracy in AI shopping and comparison answers

    If assistants consistently get basic facts wrong, check your markup, your visible content and your third-party listings for contradictions. Our guide to AI visibility metrics covers how to track accuracy over time.

    The bottom line

    Schema markup is a clarity tool. It won't make an AI assistant cite a page that doesn't answer the question, but it helps search systems understand exactly who you are and what you offer. Start with Organization markup and consistent sameAs links, add Product, Article and LocalBusiness where they fit, server-render everything, and keep it matched to what's on the page.

    FAQ

    Frequently asked questions

    Does ChatGPT read schema markup?

    OpenAI hasn't published details on how its systems use structured data. ChatGPT search draws on search indexes, and search engines like Bing and Google do use structured data to understand pages. The safest assumption is that schema helps indirectly, and the visible text on the page matters most.

    What is the most important schema for AEO?

    For most businesses, Organization schema on the homepage or about page, with name, logo, URL, description and sameAs links to official profiles. It anchors your brand as a clear entity. After that, Product for ecommerce and software, LocalBusiness for local companies, and Article for content.

    Should I still use FAQPage schema?

    Google limited FAQ rich results to well-known, authoritative government and health sites in 2023, so most sites won't get a visual result. The markup is still valid, and some other systems may use it. The bigger value is the question-and-answer content itself, which is easy for AI retrieval to match.

    JSON-LD or microdata?

    JSON-LD. Google recommends it, it's easier to maintain, and it doesn't tangle with your HTML. Make sure it's in the server-rendered HTML rather than injected late by JavaScript, so crawlers that don't run scripts can see it.

    • schema markup
    • structured data
    • JSON-LD
    • technical SEO
    • entities
    Saarthak Srivastav

    Written by

    Saarthak Srivastav

    AEO & GEO, AEO-Rank

    Saarthak Srivastav writes and reviews articles on the AEO-Rank blog, drawing on what the AEO-Rank team sees tracking AI answers across ChatGPT, Gemini, Claude, Perplexity and Grok.