TL;DR: Schema markup is the machine-readable data layer that tells AI engines what your content is, not just what it says. A controlled experiment by Search Engine Land found that only the page with well-implemented schema appeared in Google's AI Overviews — while identical pages without it were invisible. This guide covers the schema types that matter most for generative search, gives you copy-paste JSON-LD for each, and ends with a prioritised implementation checklist. It's the final article in our 7-part GEO series.

If you've followed our Generative Engine Optimisation series from the beginning, you've learned how to structure content for AI citation, clarify your entities, build co-citations, and measure what's working. This final instalment tackles the technical foundation underneath all of it: structured data.

Think of schema markup as your website's metadata passport. When a large language model retrieves your page, it doesn't just read your paragraphs — it looks for structured signals that confirm what your content is about, who created it, and whether the facts can be trusted. Without schema, you're asking AI to guess. With it, you're handing it a cheat sheet.

Traditional SEO used schema primarily to earn rich results — star ratings, recipe cards, FAQ dropdowns. That's still valuable, but the strategic importance has shifted.

AI engines use schema for grounding

Google, Microsoft, and OpenAI all confirmed in 2025 that their generative AI systems use structured data to better understand web content. As Search Engine Journal reported, schema markup builds a "content knowledge graph" — a data layer that tells machines what your brand is, what it offers, and how it should be understood.

Go Fish Digital's patent research identified Google's US9449105B1 patent for "context-aware query classification," showing that AI uses vocabulary-aware retrieval to match content to queries. When your structured data uses the same vocabulary as Schema.org, you're speaking the language these systems already understand.

The key insight from Semrush and Google's own developer documentation: your page structure, schema markup, and any commerce feeds should all describe the same thing. Consistency between what's visible on the page and what's in the structured data is critical — discrepancies can actually hurt you.

The experiment that proved it

A controlled experiment published by Search Engine Land in late 2025 tested three nearly identical pages: one with well-implemented schema, one with poor schema, and one with none. The result? Only the page with quality schema appeared in an AI Overview and achieved the highest organic ranking (Position 3). The poorly-implemented and absent-schema pages were effectively invisible to generative search.

This wasn't correlation — it was a controlled test. Schema quality, not just presence, matters.

Not all schema types carry equal weight for AI visibility. Here are the ones that directly help generative engines understand and cite your content, with JSON-LD examples you can adapt.

1. FAQPage

FAQ schema is arguably the single highest-impact type for GEO. AI engines love question-and-answer formats because they map directly to how users query these systems. When your FAQPage schema matches a user's question, you're giving the AI a pre-formatted answer it can cite.

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "What is schema markup for generative search?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Schema markup for generative search is structured data added to your web pages using Schema.org vocabulary and JSON-LD format. It helps AI engines like Google's AI Overviews and ChatGPT understand your content's meaning, entities, and relationships — making it more likely to be cited in AI-generated answers."
      }
    },
    {
      "@type": "Question",
      "name": "Does schema markup improve AI Overview visibility?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Controlled experiments show that pages with well-implemented schema are significantly more likely to appear in AI Overviews compared to identical pages without schema markup."
      }
    }
  ]
}

2. Article (with author and organisation)

Article schema tells AI engines who wrote the content, when it was published, and which organisation stands behind it. This directly supports E-E-A-T signals that generative engines use for entity clarity and trust assessment.

{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "Schema Markup and Structured Data for Generative Search",
  "author": {
    "@type": "Person",
    "name": "Jane Smith",
    "url": "https://example.com/team/jane-smith"
  },
  "publisher": {
    "@type": "Organization",
    "name": "Example Company",
    "logo": {
      "@type": "ImageObject",
      "url": "https://example.com/logo.png"
    }
  },
  "datePublished": "2026-02-18",
  "dateModified": "2026-02-18",
  "description": "A complete guide to using schema markup for AI search visibility.",
  "mainEntityOfPage": {
    "@type": "WebPage",
    "@id": "https://example.com/blog/schema-markup-guide"
  }
}

3. Organisation

Organisation schema is the foundation of entity clarity — a theme we explored in Article 4 of this series. It explicitly defines who you are, what you do, and where you operate. Without it, AI engines must infer your identity from scattered mentions across the web.

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "Hivory",
  "url": "https://hivory.io",
  "logo": "https://hivory.io/logo.png",
  "description": "Hivory builds AI-powered analytics tools including Meaning, a Google Analytics chatbot.",
  "sameAs": [
    "https://www.linkedin.com/company/hivory",
    "https://twitter.com/hivory"
  ],
  "foundingDate": "2023",
  "contactPoint": {
    "@type": "ContactPoint",
    "contactType": "customer service",
    "url": "https://hivory.io/contact"
  }
}

The sameAs property is particularly valuable — it links your entity to your social profiles, helping AI engines build a complete picture of your brand through co-citations and brand mentions.

4. HowTo

HowTo schema structures procedural content in a way AI engines can directly parse and present. If your content explains processes — installation guides, setup instructions, marketing workflows — this schema type makes it far more citable.

{
  "@context": "https://schema.org",
  "@type": "HowTo",
  "name": "How to Add Schema Markup to Your Website",
  "description": "A step-by-step guide to implementing JSON-LD schema markup for improved AI search visibility.",
  "step": [
    {
      "@type": "HowToStep",
      "name": "Audit your existing schema",
      "text": "Use Google's Rich Results Test to check what structured data your pages currently have. Identify gaps in entity coverage."
    },
    {
      "@type": "HowToStep",
      "name": "Prioritise high-impact schema types",
      "text": "Start with Organisation and Article schema on your key pages, then add FAQPage schema to content that answers common questions."
    },
    {
      "@type": "HowToStep",
      "name": "Implement using JSON-LD",
      "text": "Add JSON-LD script tags to your page head or body. JSON-LD is Google's recommended format as it separates structured data from HTML markup."
    },
    {
      "@type": "HowToStep",
      "name": "Validate and monitor",
      "text": "Test with Schema Markup Validator and Google Rich Results Test. Monitor for errors in Google Search Console."
    }
  ]
}

5. other high-value types

Beyond the four above, several other schema types are worth implementing depending on your content:

  • Product — Essential for e-commerce. Include name, description, price, availability, review, and aggregateRating. Products with comprehensive schema are up to 3x more likely to appear in AI-generated shopping recommendations.
  • BreadcrumbList — Helps AI engines understand your site's hierarchy and the relationships between pages. Simple to implement, often overlooked.
  • Dataset — If you publish original research, data, or statistics, Dataset schema makes your data directly discoverable by AI engines looking for factual grounding.
  • SpeakableSpecification — Marks content sections as suitable for text-to-speech playback. As voice-based AI assistants grow, this schema type signals which parts of your content are best suited for spoken answers.

JSON-LD best practices for GEO

JSON-LD (JavaScript Object Notation for Linked Data) is the format Google recommends for structured data, and it's what AI engines parse most reliably. Here are the practices that matter:

Do

  • Place JSON-LD in your <head> or <body> — either works, but <head> is cleaner
  • Use @id references to connect related schema objects (e.g., link your Article's author to a Person entity, and that Person to your Organisation)
  • Match your visible content — if your schema says "Product: Blue Widget, £49.99" but the page shows "Azure Gadget, £59.99," you've created a trust problem
  • Nest entities rather than creating disconnected schema blocks — this builds the relationship graph AI engines need
  • Keep it current — update dateModified when content changes; stale dates signal neglect

Don't

  • Don't mark up content that isn't visible on the page — Google explicitly penalises this
  • Don't use Microdata or RDFa for new implementations — JSON-LD is easier to maintain and doesn't tangle with your HTML
  • Don't over-schema — marking every paragraph as a "Thing" adds noise, not signal
  • Don't forget validation — a single syntax error can invalidate your entire schema block

Testing and validation tools

Before you deploy schema to production, validate it:

  1. Google Rich Results Test — Tests whether your schema qualifies for rich results and catches errors
  2. Schema Markup Validator — The official Schema.org validator; checks syntax and vocabulary compliance
  3. Google Search Console — Monitor structured data errors and warnings across your entire site over time
  4. Semrush Site Audit — Includes structured data checks alongside broader SEO auditing

After implementing schema, use Meaning to track whether your structured pages see improved engagement or AI referral traffic. Ask Meaning: "Compare traffic on pages with schema markup vs pages without it over the last 30 days."

Prioritised implementation checklist

Not sure where to start? Here's the order that delivers the most impact for the least effort:

Priority 1: foundation (week 1)

  • Add Organisation schema to your homepage
  • Add Article schema to all blog posts and content pages
  • Validate all existing schema with Google Rich Results Test
  • Fix any schema errors flagged in Google Search Console

Priority 2: quick wins (week 2)

  • Add FAQPage schema to pages that answer common questions
  • Add BreadcrumbList schema across your site
  • Ensure all schema matches visible page content exactly
  • Connect entities using @id references

Priority 3: advanced (week 3-4)

  • Add HowTo schema to procedural/tutorial content
  • Add Product schema to product and pricing pages
  • Implement SpeakableSpecification on key content
  • Add Dataset schema to any original research or data pages
  • Build a connected entity graph across your site using nested schema

Priority 4: monitor and iterate (ongoing)

  • Set up weekly schema error monitoring in Search Console
  • Track AI referral traffic to schema-enhanced pages in Meaning
  • Update dateModified whenever content changes
  • A/B test schema implementations on similar pages to measure impact

Common mistakes to avoid

Inconsistency between schema and visible content. This is the number one mistake. If your schema describes a different product, price, or service than what's on the page, AI engines will distrust both signals.

Implementing schema without a strategy. Adding random schema types to random pages creates noise. Start with your most important pages and the schema types that match your content.

Forgetting to update. Schema with a datePublished of 2022 and no dateModified tells AI engines your content may be stale — exactly the kind of signal that keeps you out of AI Overviews.

Ignoring the knowledge graph. Individual schema blocks help, but connected schema — where your Organisation links to your Articles, which link to their Authors, who link back to the Organisation — creates the relationship graph that AI engines use for entity understanding. As we covered in our article on content structure for AI citation, structure is everything.

Frequently asked questions

Does schema markup directly improve AI Overview rankings?

Controlled experiments show a strong correlation. In Search Engine Land's 2025 test, only the page with well-implemented schema appeared in AI Overviews. While Google hasn't confirmed schema as a direct ranking factor for AI Overviews, the evidence suggests it plays a meaningful role in helping AI engines understand and select content for citation.

Which schema type should I implement first?

Start with Organisation schema on your homepage and Article schema on your content pages. These establish the foundational entities — who you are and what you've published. Then add FAQPage schema to your most question-focused content for quick GEO wins.

Is JSON-LD the only format I should use?

Yes, for new implementations. JSON-LD is Google's recommended format, it's easier to maintain than Microdata or RDFa, and it keeps your structured data separate from your HTML. All major AI systems parse JSON-LD reliably.

Use Google Search Console to monitor for schema errors, and use Meaning to track engagement and traffic patterns on your schema-enhanced pages. Ask Meaning to compare traffic on pages with schema vs those without — look for increases in AI referral traffic and engagement metrics. We covered this in depth in Article 6: Tracking Your GEO Performance.

Can schema markup hurt my site if done incorrectly?

Yes. Poorly implemented schema — especially schema that contradicts your visible content — can reduce trust signals. Google may also issue manual actions for deliberately misleading structured data. Always validate before deploying.

How often should I update my schema?

Update your schema whenever the underlying content changes. At minimum, update dateModified with each content revision. Review your site's schema quarterly to catch errors and add coverage to new pages.

Series wrap-up: the complete GEO playbook

This article completes our 7-part series on Generative Engine Optimisation. Here's what we've covered — and how it all connects:

  1. What Is GEO — The landscape shift from traditional SEO to optimising for AI-generated answers
  2. 9 Proven Techniques That Boost AI Search Visibility by 40% — The research-backed tactics that move the needle
  3. How to Structure Content So AI Engines Actually Cite You — Formatting, hierarchy, and passage-level optimisation
  4. Entity Clarity: Why AI Can't Recommend You If It Doesn't Understand What You Do — Making your brand unambiguous to machines
  5. Co-Citations and Brand Mentions: The New Backlinks for AI Search — Building authority through association
  6. Tracking Your GEO Performance: Metrics That Actually Matter — Measuring what works in the AI search era
  7. Schema Markup and Structured Data for Generative Search (this article) — The technical foundation that ties it all together

The through-line across all seven articles is this: AI engines reward clarity. Clear content structure, clear entity definitions, clear authority signals, and clear structured data. If you've implemented the strategies across this series, you've built a comprehensive GEO foundation.

The next step? Measure everything. Connect Meaning to your Google Analytics, ask it questions about your AI search performance, and let the data tell you where to focus next. Generative search isn't a one-time optimisation — it's an ongoing conversation between your content and the AI engines that surface it.

The brands that win in AI search won't be the ones with the biggest budgets. They'll be the ones that made themselves the easiest to understand.