Structured Data for LLMs: The Schema Markup Guide That Gets You Cited by AI
71% of pages cited by ChatGPT include schema markup. Pages with proper structured data are cited 3.2x more often in AI-generated responses. This guide gives you the 5 essential JSON-LD schema types — with complete, production-ready code examples — that get your brand cited across ChatGPT, Perplexity, and Google AI Overviews.
What is schema markup — and why does it drive AI citations?
Structured data markup — specifically JSON-LD schema — is the machine-readable layer that tells AI platforms exactly what your content means, who created it, and why it is trustworthy. Pages with proper schema markup are cited 3.2x more often in AI-generated responses than pages without it. The five schema types that drive the most AI citations are: Organization, Article, FAQPage, HowTo, and Person.
Most SEO guides treat schema as a way to earn rich results in Google — star ratings, FAQ dropdowns, price displays. That is still true. But in 2026, schema has become something more fundamental: it is the primary language that AI systems use to read, trust, and cite your content.
Why schema markup specifically helps LLMs — the technical reason
When an AI crawler retrieves your page, it reads raw HTML. This means it sees your content as unstructured text — words and sentences without inherent meaning. The AI must then infer what everything means: is "Kongzilla" a brand or a product? Is "Somesh Tripathi" an author, a client, or a subject? Is "April 2026" a publication date or a reference to an event?
Schema markup removes all of that inferencing. By adding JSON-LD code to your pages, you explicitly tell the AI exactly what everything is — in a language it was designed to read.
AI must infer: Is this a brand? An author? A date? A product name? Gets it wrong, or skips the page entirely for a clearer source.
"@type": "Person", "name": "Somesh Tripathi"
"dateModified": "2026-04-10"
No guessing. No ambiguity. Direct citation signal.
Think of schema markup as the difference between handing an AI a pile of raw ingredients versus handing it a labelled recipe card. Both contain the same information. But only one tells the AI exactly what each ingredient is, how much of it there is, and what to do with it. Schema markup is the recipe card. Without it, AI systems either guess incorrectly or skip your content entirely in favour of a source that communicates more clearly.
The three tiers of schema types — implement in this order
Not all schema types carry equal weight for AI citations. Implement in tier order for the fastest return on effort.
The 5 essential schema types — with complete JSON-LD code examples
Production-ready code for each schema type. Copy, update with your details, and paste into your page's <head> in a <script type="application/ld+json"> block.
Organization schema tells every AI system who you are. Without it, LLMs piece together your brand identity from scattered web mentions — and frequently get it wrong. The sameAs property is the most important field: it links your website to verified profiles on LinkedIn, Crunchbase, and Wikipedia, creating an entity graph AI systems can cross-reference.
Pages with connected sameAs references receive 3–5x higher citation rates than pages with isolated schema blocks (ADV Strategy Pro, 2026).
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Organization",
"@id": "https://kongzilla.co/#organization",
"name": "Kongzilla",
"url": "https://kongzilla.co",
"logo": "https://kongzilla.co/logo.png",
"description": "AI SEO agency based in India specialising in GEO,
AEO, and AI search optimisation for businesses across
the UK, Australia, and India.",
"foundingDate": "2018",
"address": {
"@type": "PostalAddress",
"addressLocality": "Kharar",
"addressRegion": "Punjab",
"addressCountry": "IN"
},
"sameAs": [
"https://www.linkedin.com/company/kongzilla",
"https://www.crunchbase.com/organization/kongzilla",
"https://g2.com/sellers/kongzilla"
],
"knowsAbout": [
"Generative Engine Optimisation",
"Answer Engine Optimisation",
"AI SEO",
"White-label SEO",
"GEO optimization"
]
}
</script>
The two most critical properties are dateModified (freshness signal) and author (E-E-A-T signal). AI systems have a strong recency bias — 95% of ChatGPT citations come from content updated within 10 months — and dateModified is how you communicate that signal explicitly.
Notice the "@id": "https://kongzilla.co/#organization" in the publisher field. This connects this Article to your Organization schema — creating a linked entity graph that AI systems recognise as a coherent, trustworthy source.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Article",
"headline": "What is Generative Engine Optimisation (GEO)?
The Complete 2026 Guide",
"description": "GEO is the practice of optimising content so AI
platforms cite your brand in generated answers.",
"datePublished": "2026-04-01",
"dateModified": "2026-04-10",
"author": {
"@type": "Person",
"name": "Somesh Tripathi",
"url": "https://www.linkedin.com/in/someshtripathi",
"jobTitle": "Founder & CEO, Kongzilla"
},
"publisher": {
"@type": "Organization",
"@id": "https://kongzilla.co/#organization",
"name": "Kongzilla"
},
"mainEntityOfPage": {
"@type": "WebPage",
"@id": "https://kongzilla.co/blog/what-is-geo/"
}
}
</script>
FAQPage schema is the single most impactful schema type for AI citation rates. LLMs are fundamentally question-answering machines, and FAQPage schema maps directly to how they retrieve and present information. Each FAQ entry is an independent citation candidate — a self-contained question and answer that AI can extract for a different query.
Three rules for maximum citation probability:
- Keep each answer to 40–80 words and fully self-contained — an AI must be able to use the answer without surrounding context
- Make the question text exactly match what you want AI to answer
- Never mark up content with FAQPage schema unless those questions visibly appear on the page — schema that does not match visible content is flagged as deceptive and penalised
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [
{
"@type": "Question",
"name": "What is generative engine optimisation?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Generative engine optimisation (GEO) is the practice
of structuring your content so AI platforms like ChatGPT,
Perplexity, and Google AI Overviews cite your brand when
answering user questions. It differs from traditional SEO
by targeting AI-generated answers rather than ranked links."
}
},
{
"@type": "Question",
"name": "How long does GEO take to show results?",
"acceptedAnswer": {
"@type": "Answer",
"text": "GEO optimisation typically produces first citation results
within 2–6 weeks for niche queries. More competitive topics
require 3–6 months. Perplexity shows results fastest;
ChatGPT's base model takes 6–18 months to fully reflect changes."
}
}
]
}
</script>
HowTo schema maps step-by-step processes into a machine-readable format that AI systems extract directly for procedural queries. Any content structured as numbered steps — "how to allow ChatGPT to crawl your website", "how to implement FAQ schema", "how to set up llms.txt" — should have HowTo schema applied.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "HowTo",
"name": "How to Allow ChatGPT to Crawl Your Website",
"description": "A step-by-step guide to configuring your robots.txt
to allow GPTBot and OAI-SearchBot access to your content
for ChatGPT citations.",
"totalTime": "PT15M",
"step": [
{
"@type": "HowToStep",
"name": "Open your robots.txt file",
"text": "Navigate to yourdomain.com/robots.txt in your browser
to see its current contents. Look for any lines that
Disallow GPTBot or OAI-SearchBot."
},
{
"@type": "HowToStep",
"name": "Add Allow directives for OpenAI crawlers",
"text": "Add 'User-agent: GPTBot Allow: /' and
'User-agent: OAI-SearchBot Allow: /' to your robots.txt.
Save and re-upload."
},
{
"@type": "HowToStep",
"name": "Check Cloudflare Bot Fight Mode",
"text": "If using Cloudflare, navigate to Security › Bots and
verify Bot Fight Mode is not blocking AI crawlers at the
network level before your robots.txt is even read."
}
]
}
</script>
Person schema establishes author identity — one of the most significant E-E-A-T signals AI systems evaluate. AI systems treat clearly attributed, credentialed content as more trustworthy than anonymous content, exactly mirroring Google's own guidance on Experience, Expertise, Authoritativeness, and Trustworthiness.
Notice the "@id" on the Person and the "worksFor" linking back to the Organization @id. This is the nesting technique that multiplies citation rates — covered in detail in the next section.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Person",
"@id": "https://kongzilla.co/#somesh-tripathi",
"name": "Somesh Tripathi",
"jobTitle": "Founder & CEO",
"worksFor": {
"@type": "Organization",
"@id": "https://kongzilla.co/#organization"
},
"url": "https://www.linkedin.com/in/someshtripathi",
"knowsAbout": [
"AI SEO",
"Generative Engine Optimisation",
"Digital Marketing",
"Content Strategy"
]
}
</script>
Schema nesting — the @id technique that multiplies citation rates by 3–5x
Individual schema blocks are useful. Nested schema blocks — where related schemas are connected through shared @id references — are dramatically more powerful for AI citations.
When you link your Article schema to your Organization schema through a shared @id, and link your Person schema to your Organization schema through worksFor, you create what researchers call a content knowledge graph. AI systems can follow these relationships to understand that this Article was written by this Person who works for this Organisation that knowsAbout these topics.
knowsAbout: GEO, AEO, AI SEO
worksFor → Organization @id
publisher → Organization @id
- Every Article schema → references the same Organisation @id in its
publisherfield - Every Person schema → references the same Organisation @id in its
worksForfield - Every Service page → references the Organisation @id in its
providerfield - The @id value is always the same:
https://kongzilla.co/#organization
The 6 schema mistakes that silently block AI citations
Each of these errors can invalidate your entire schema implementation — with no error visible to you, but zero benefit in AI citations.
<script type="application/ld+json"> block in the page <head>.Complete schema implementation checklist — minimum viable stack for AI citations
This checklist covers the minimum viable schema stack for consistent AI citations across ChatGPT, Perplexity, and Google AI Overviews.
- Organization schema in JSON-LD with: @id, name, url, logo, description, foundingDate, address, sameAs (LinkedIn, Crunchbase, G2), knowsAbout
- WebSite schema with SearchAction (enables sitelinks search box in Google)
- Validated in Google Rich Results Test with zero errors
- Page source confirms only ONE Organization block exists — no plugin duplicates
- Article schema with: headline, description, datePublished, dateModified, author (Person @id), publisher (Organization @id), mainEntityOfPage
- FAQPage schema if the post contains a visible FAQ section (minimum 3 questions)
- HowTo schema on any numbered step-by-step process section
- Person schema on author bio section, linked to Organization via worksFor @id
- dateModified updated every time the content is refreshed — not just datePublished
- Service schema with: name, description, provider (Organization @id), areaServed, url
- FAQPage schema if the service page includes visible questions and answers
- BreadcrumbList schema on all pages below homepage (helps AI understand site structure)
- Google Search Console Enhancements report checked monthly for schema errors
- dateModified updated on all high-priority pages every 6–8 weeks
- New pages validated in Rich Results Test before publishing
- sameAs links in Organization schema checked quarterly — remove any that are outdated or broken
Tools to implement and validate schema markup
You do not need to hand-code every JSON-LD block from scratch. These tools make implementation faster and validation reliable.
Frequently Asked Questions
Which schema gaps are costing you AI citations today?
Kongzilla is an AI SEO agency based in India specialising in GEO, AEO, and complete AI search optimisation for businesses across the UK, Australia, and India. We implement the full technical schema stack — Organization, Article, FAQPage, HowTo, Person, and Service schema — as part of every client engagement, alongside content strategy and entity authority building. Book a free AI visibility audit and see exactly which schema gaps are costing you citations today.
