{ "@context": "https://agentflare.org/schema", "type": "Article", "tier": "L2-full", "title": "Structured Data for LLMs: JSON-LD That Agents Actually Use", "description": "Schema.org JSON-LD is a low-cost way to label articles, FAQs, how-tos and entities for search engines and LLM retrieval. Evidence of use is real but modest,…", "canonical": "https://agentflare.org/research/structured-data-for-llms-json-ld-that-agents-actually-use.html", "category": "research", "updated": "2026-10-16", "generated_at": "2026-10-16T00:37:01.652Z", "facts": [ { "label": "Cluster", "value": "Discovery & crawling" }, { "label": "Sources", "value": "8" }, { "label": "Updated", "value": "2026-10-16" } ], "data": { "topic": "structured data JSON-LD schema.org for LLM consumption", "cluster": "discovery", "generated_at": "2026-10-16T00:37:01.652Z", "tldr": "Schema.org JSON-LD is a low-cost way to label articles, FAQs, how-tos and entities for search engines and LLM retrieval. Evidence of use is real but modest, so deploy a small validated set by template and keep it matched to page content.", "intro": "JSON-LD is the embedded format for Schema.org markup, a vocabulary launched in 2011 by Google, Bing, Yahoo and Yandex. Sources describe it as a machine-readable layer that ChatGPT, Claude, Perplexity and Gemini can consume to classify content, identify authors and extract citable claims. Claims about its influence vary. Microsoft Bing has confirmed it uses schema for its LLMs, one study finds schema on around 65% of pages cited by AI Mode, and Google says structured data does not directly influence rankings. This brief covers which types to deploy, how to map them to templates, how to validate them, and where the evidence is weak, so developers and agents can implement it with realistic expectations.", "concepts": [ { "term": "Schema.org", "definition": "An open vocabulary of machine-readable types such as Article, FAQPage, HowTo, Organization, Person and DefinedTerm, launched in 2011 by Google, Microsoft Bing, Yahoo and Yandex.", "sources": [ "S1", "S4" ] }, { "term": "JSON-LD", "definition": "JavaScript Object Notation for Linked Data: a lightweight text format that conveys entities and their relationships and can be added to a page without altering its HTML.", "sources": [ "S2" ] }, { "term": "FAQPage markup", "definition": "Question entities with a name and an acceptedAnswer of type Answer with text. Self-contained answers give the highest extraction rate for AI citation, according to S1.", "sources": [ "S1" ] }, { "term": "Shared Organization @id", "definition": "Referencing the same Organization @id across all pages lets LLMs disambiguate the publisher entity, which is defined with name, url, logo and sameAs.", "sources": [ "S1" ] }, { "term": "Generative Engine Optimization", "definition": "GEO is the practice of influencing how AI systems interpret and reference a brand, with schema markup as one supporting method.", "sources": [ "S4", "S8" ] } ], "how_it_works": { "title": "Deploying JSON-LD across a site", "steps": [ { "step": "Audit existing markup", "detail": "Run Google's Rich Results Test on 10-20 representative URLs, including homepage, pillar pages, articles, product pages and the About page. Note which types are present, missing or malformed, and build on CMS-generated markup." }, { "step": "Map types to templates", "detail": "Assign schema types per page template, not per URL: Article for posts, DefinedTerm for glossary entries, HowTo for guides, Organization plus Person for About pages. Choose the smallest set covering 95% of content." }, { "step": "Write the JSON-LD fields", "detail": "Include required and recommended properties. For Article: headline, author, publisher, datePublished, dateModified, image, mainEntityOfPage. Reuse one Organization @id site-wide." }, { "step": "Validate the markup", "detail": "Check with Google Rich Results Test, validator.schema.org, Bing Markup Validator and Search Console coverage reports. Confirm schema content, such as HowTo step counts, matches what appears on the page." }, { "step": "Monitor and maintain", "detail": "Review Search Console structured data reports after deployment and keep dateModified current, since freshness signals are weighted heavily by LLM retrieval systems according to S1." } ] }, "sections": [ { "h": "Schema types worth deploying", "body": "Map schema types to page templates rather than individual URLs. S1 recommends the smallest set that covers 95% of content, typically Article, FAQPage, HowTo, Organization, Person, DefinedTerm and BreadcrumbList. Article needs headline, author as Person, publisher as Organization, datePublished, dateModified, image and mainEntityOfPage. Both dates should always be present because freshness signals are weighted heavily by LLM retrieval systems. FAQPage entries use a Question with a name and an acceptedAnswer of type Answer with text. Answers of 2-4 sentences stay quotable in isolation, and five to eight per pillar page is described as the practical sweet spot. HowTo needs HowToStep entities with name and text, and step counts must match the page. Organization with name, url, logo and sameAs builds the publisher entity, and reusing the same Organization @id across pages helps LLMs disambiguate it. Person markup on author bios ties each article to a real human with credentials.", "sources": [ "S1", "S4" ] }, { "h": "What the evidence says about LLM use", "body": "S7 reports that in March 2025 Fabrice Canel of Microsoft Bing confirmed at SMX Munich that schema markup helps Microsoft's LLMs understand content, specifically for Copilot. S6 found that around 65% of pages cited by AI Mode and 71% cited by ChatGPT use schema markup. The most common types on AI-cited pages were BreadcrumbList (38–42%), WebSite (34–36%), Organization (31–32%) and WebPage (31–36%). Most AI-cited pages use just 1–2 types (43% for AI Mode, 42% for ChatGPT). S6 also reports that structural schemas such as BreadcrumbList or WebPage show weak negative correlations with AI citations, while Article shows a weak positive one. It calls the effects tiny and says clear, well-structured content matters most. S8 notes John Mueller of Google confirmed in 2025 that structured data does not directly influence rankings. S2 and S3 argue more strongly that LLMs absorb JSON-LD as a data layer, a view S3 ties to reasoning models, retrieval architectures and knowledge graphs.", "sources": [ "S2", "S3", "S6", "S7", "S8" ] }, { "h": "Validation and LLM-generated markup", "body": "Validate before and after deployment. S1 lists Google Rich Results Test, validator.schema.org and Search Console structured data coverage reports, and S6 adds Bing Markup Validator. Keep schema consistent with visible page content, and avoid markup such as third-party reviews or inactive job postings, which S6 says can put SEO and AI visibility at risk. Generating markup with LLMs needs the same care. S5 reports that around 75% of web pages still lack Schema.org markup, and that 40%–50% of markup produced by GPT-3.5 and GPT-4 is invalid, non-factual or non-compliant with the Schema.org ontology, which has 806 types and over 1,400 properties. Specialized LLM-powered agents that filter these errors improved quality, and after filtering GPT-4 outperformed humans. For agents that write JSON-LD, a validation and filtering step is therefore part of the workflow, not an optional extra.", "sources": [ "S1", "S5", "S6" ] } ], "comparison": { "title": "Classical SEO schema vs LLM-oriented JSON-LD", "columns": [ "Aspect", "Classical SEO schema", "LLM-oriented JSON-LD" ], "rows": [ [ "Supported types", "~30 officially supported", "Any type plus custom vocabularies" ], [ "Schema density", "Minimal, just enough", "Maximal, be verbose" ], [ "Main goal", "Prettier results and higher CTR", "Helping models understand and remember entities" ], [ "Tolerance", "Very low; unsupported markup ignored", "Models weigh meaning over syntax, per S2" ] ], "sources": [ "S2" ] }, "key_numbers": [ { "value": "71%", "label": "of ChatGPT-cited pages use schema markup", "source": "S6" }, { "value": "65%", "label": "of AI Mode-cited pages use schema markup", "source": "S6" }, { "value": "40%–50%", "label": "of GPT-3.5 and GPT-4 markup invalid or non-compliant", "source": "S5" }, { "value": "75%", "label": "of web pages still lack Schema.org markup", "source": "S5" } ], "timeline": [ { "date": "2011", "event": "Schema.org launched by Google, Bing, Yahoo and Yandex", "source": "S4" }, { "date": "2025-03", "event": "Bing's Fabrice Canel confirms schema helps Microsoft LLMs at SMX Munich", "source": "S7" }, { "date": "2025", "event": "Google's John Mueller says structured data does not directly influence rankings", "source": "S8" }, { "date": "2025", "event": "LLM4Schema.org benchmark published, assessing LLM-generated Schema.org markup", "source": "S5" } ], "pitfalls": [ { "text": "Schema that does not match page content, such as HowTo step counts differing from the visible steps, is a common Rich Results Test failure.", "source": "S1" }, { "text": "LLM-written markup is unreliable unfiltered: 40%–50% of GPT-3.5 and GPT-4 output was invalid, non-factual or non-compliant with the ontology.", "source": "S5" }, { "text": "Do not over-credit schema: structural types showed weak negative correlations with AI citations, and S6 calls the effects tiny.", "source": "S6" }, { "text": "Markup for third-party reviews or inactive job postings can put SEO and AI visibility at risk.", "source": "S6" } ], "takeaways": [ "Map a small set of validated schema types to page templates, and keep datePublished, dateModified and a shared Organization @id consistent.", "Evidence of LLM use is confirmed for Bing and correlational elsewhere; Google says rankings are not directly affected.", "Validate all markup, especially anything generated by an LLM, and keep it consistent with visible page content." ], "faq": [ { "q": "Which schema types should I add first for LLM visibility?", "a": "S1 suggests Article, FAQPage, HowTo, Organization, Person, DefinedTerm and BreadcrumbList, mapped to templates. S4 highlights FAQPage, Article, Organization, Person and WebPage. Start with Article and Organization, then add FAQPage where answers are self-contained." }, { "q": "Do LLMs actually read JSON-LD?", "a": "Microsoft Bing's Fabrice Canel confirmed in March 2025 that schema helps its LLMs understand content. S6 found schema on around 65% of AI Mode-cited pages and 71% of ChatGPT-cited pages. Other platforms have no equivalent confirmation in these sources." }, { "q": "Does JSON-LD improve Google rankings?", "a": "S8 reports that John Mueller confirmed in 2025 that structured data does not directly influence rankings. It still supports rich results and user engagement, and helps search systems classify content." }, { "q": "Can I let an LLM generate my JSON-LD?", "a": "Only with validation. S5 found 40%–50% of GPT-3.5 and GPT-4 markup was invalid, non-factual or non-compliant. Filtering by specialized LLM agents improved quality, and GPT-4 then outperformed humans. Run validator.schema.org and the Rich Results Test on output." } ], "sources": [ { "title": "Schema for AI: Schema.org & JSON-LD for LLM Citations (2026)", "url": "https://alicelabs.ai/en/insights/schema-org-for-ai", "domain": "alicelabs.ai", "date": "" }, { "title": "Schema & JSON-LD for LLM Search: Structured Data for LLMO / AEO / GEO | Szymon Słowik", "url": "https://www.szymonslowik.com/json-ld-for-llm-seo", "domain": "szymonslowik.com", "date": "" }, { "title": "Structured Data, Not Tokenization, is the Future of LLMs", "url": "https://www.schemaapp.com/schema-markup/why-structured-data-not-tokenization-is-the-future-of-llms", "domain": "schemaapp.com", "date": "" }, { "title": "How Can Schema Markup Specifically Enhance LLM Visibility | Walker Sands", "url": "https://www.walkersands.com/about/blog/how-can-schema-markup-support-llm-visibility", "domain": "walkersands.com", "date": "" }, { "title": "LLM4Schema.org: Generating Schema.org Markups With Large Language Models - Minh-Hoang Dang, Thi Hoang Thi Pham, Pascal Molli, Hala Skaf-Molli, Alban Gaignard, 2025", "url": "https://journals.sagepub.com/doi/10.1177/22104968251382172", "domain": "journals.sagepub.com", "date": "" }, { "title": "Structured Data for SEO and LLMs | Schema.org Markup Guide", "url": "https://seranking.com/blog/structured-data", "domain": "seranking.com", "date": "" }, { "title": "Schema & Structured Data for LLM Visibility: What Actually Helps?", "url": "https://www.quoleady.com/schema-structured-data-for-llm-visibility", "domain": "quoleady.com", "date": "" }, { "title": "Structured data: SEO and GEO optimization for AI in 2026", "url": "https://www.digidop.com/blog/structured-data-secret-weapon-seo", "domain": "digidop.com", "date": "" } ] }, "analysis_md": "Schema.org JSON-LD is a low-cost way to label articles, FAQs, how-tos and entities for search engines and LLM retrieval. Evidence of use is real but modest, so deploy a small validated set by template and keep it matched to page content.\n\nJSON-LD is the embedded format for Schema.org markup, a vocabulary launched in 2011 by Google, Bing, Yahoo and Yandex. Sources describe it as a machine-readable layer that ChatGPT, Claude, Perplexity and Gemini can consume to classify content, identify authors and extract citable claims. Claims about its influence vary. Microsoft Bing has confirmed it uses schema for its LLMs, one study finds schema on around 65% of pages cited by AI Mode, and Google says structured data does not directly influence rankings. This brief covers which types to deploy, how to map them to templates, how to validate them, and where the evidence is weak, so developers and agents can implement it with realistic expectations.\n\n## Key numbers\n\n- **71%** — of ChatGPT-cited pages use schema markup (https://seranking.com/blog/structured-data)\n- **65%** — of AI Mode-cited pages use schema markup (https://seranking.com/blog/structured-data)\n- **40%–50%** — of GPT-3.5 and GPT-4 markup invalid or non-compliant (https://journals.sagepub.com/doi/10.1177/22104968251382172)\n- **75%** — of web pages still lack Schema.org markup (https://journals.sagepub.com/doi/10.1177/22104968251382172)\n\n## Key concepts\n\n- **Schema.org:** An open vocabulary of machine-readable types such as Article, FAQPage, HowTo, Organization, Person and DefinedTerm, launched in 2011 by Google, Microsoft Bing, Yahoo and Yandex.\n- **JSON-LD:** JavaScript Object Notation for Linked Data: a lightweight text format that conveys entities and their relationships and can be added to a page without altering its HTML.\n- **FAQPage markup:** Question entities with a name and an acceptedAnswer of type Answer with text. Self-contained answers give the highest extraction rate for AI citation, according to S1.\n- **Shared Organization @id:** Referencing the same Organization @id across all pages lets LLMs disambiguate the publisher entity, which is defined with name, url, logo and sameAs.\n- **Generative Engine Optimization:** GEO is the practice of influencing how AI systems interpret and reference a brand, with schema markup as one supporting method.\n\n## Deploying JSON-LD across a site\n\n1. **Audit existing markup** — Run Google's Rich Results Test on 10-20 representative URLs, including homepage, pillar pages, articles, product pages and the About page. Note which types are present, missing or malformed, and build on CMS-generated markup.\n2. **Map types to templates** — Assign schema types per page template, not per URL: Article for posts, DefinedTerm for glossary entries, HowTo for guides, Organization plus Person for About pages. Choose the smallest set covering 95% of content.\n3. **Write the JSON-LD fields** — Include required and recommended properties. For Article: headline, author, publisher, datePublished, dateModified, image, mainEntityOfPage. Reuse one Organization @id site-wide.\n4. **Validate the markup** — Check with Google Rich Results Test, validator.schema.org, Bing Markup Validator and Search Console coverage reports. Confirm schema content, such as HowTo step counts, matches what appears on the page.\n5. **Monitor and maintain** — Review Search Console structured data reports after deployment and keep dateModified current, since freshness signals are weighted heavily by LLM retrieval systems according to S1.\n\n## Schema types worth deploying\n\nMap schema types to page templates rather than individual URLs. S1 recommends the smallest set that covers 95% of content, typically Article, FAQPage, HowTo, Organization, Person, DefinedTerm and BreadcrumbList. Article needs headline, author as Person, publisher as Organization, datePublished, dateModified, image and mainEntityOfPage. Both dates should always be present because freshness signals are weighted heavily by LLM retrieval systems. FAQPage entries use a Question with a name and an acceptedAnswer of type Answer with text. Answers of 2-4 sentences stay quotable in isolation, and five to eight per pillar page is described as the practical sweet spot. HowTo needs HowToStep entities with name and text, and step counts must match the page. Organization with name, url, logo and sameAs builds the publisher entity, and reusing the same Organization @id across pages helps LLMs disambiguate it. Person markup on author bios ties each article to a real human with credentials.\n\n## What the evidence says about LLM use\n\nS7 reports that in March 2025 Fabrice Canel of Microsoft Bing confirmed at SMX Munich that schema markup helps Microsoft's LLMs understand content, specifically for Copilot. S6 found that around 65% of pages cited by AI Mode and 71% cited by ChatGPT use schema markup. The most common types on AI-cited pages were BreadcrumbList (38–42%), WebSite (34–36%), Organization (31–32%) and WebPage (31–36%). Most AI-cited pages use just 1–2 types (43% for AI Mode, 42% for ChatGPT). S6 also reports that structural schemas such as BreadcrumbList or WebPage show weak negative correlations with AI citations, while Article shows a weak positive one. It calls the effects tiny and says clear, well-structured content matters most. S8 notes John Mueller of Google confirmed in 2025 that structured data does not directly influence rankings. S2 and S3 argue more strongly that LLMs absorb JSON-LD as a data layer, a view S3 ties to reasoning models, retrieval architectures and knowledge graphs.\n\n## Validation and LLM-generated markup\n\nValidate before and after deployment. S1 lists Google Rich Results Test, validator.schema.org and Search Console structured data coverage reports, and S6 adds Bing Markup Validator. Keep schema consistent with visible page content, and avoid markup such as third-party reviews or inactive job postings, which S6 says can put SEO and AI visibility at risk. Generating markup with LLMs needs the same care. S5 reports that around 75% of web pages still lack Schema.org markup, and that 40%–50% of markup produced by GPT-3.5 and GPT-4 is invalid, non-factual or non-compliant with the Schema.org ontology, which has 806 types and over 1,400 properties. Specialized LLM-powered agents that filter these errors improved quality, and after filtering GPT-4 outperformed humans. For agents that write JSON-LD, a validation and filtering step is therefore part of the workflow, not an optional extra.\n\n## Classical SEO schema vs LLM-oriented JSON-LD\n\n| Aspect | Classical SEO schema | LLM-oriented JSON-LD |\n| --- | --- | --- |\n| Supported types | ~30 officially supported | Any type plus custom vocabularies |\n| Schema density | Minimal, just enough | Maximal, be verbose |\n| Main goal | Prettier results and higher CTR | Helping models understand and remember entities |\n| Tolerance | Very low; unsupported markup ignored | Models weigh meaning over syntax, per S2 |\n\n## Timeline\n\n- 2011: Schema.org launched by Google, Bing, Yahoo and Yandex\n- 2025-03: Bing's Fabrice Canel confirms schema helps Microsoft LLMs at SMX Munich\n- 2025: Google's John Mueller says structured data does not directly influence rankings\n- 2025: LLM4Schema.org benchmark published, assessing LLM-generated Schema.org markup\n\n## Common pitfalls\n\n- Schema that does not match page content, such as HowTo step counts differing from the visible steps, is a common Rich Results Test failure.\n- LLM-written markup is unreliable unfiltered: 40%–50% of GPT-3.5 and GPT-4 output was invalid, non-factual or non-compliant with the ontology.\n- Do not over-credit schema: structural types showed weak negative correlations with AI citations, and S6 calls the effects tiny.\n- Markup for third-party reviews or inactive job postings can put SEO and AI visibility at risk.\n\n## Key takeaways\n\n- Map a small set of validated schema types to page templates, and keep datePublished, dateModified and a shared Organization @id consistent.\n- Evidence of LLM use is confirmed for Bing and correlational elsewhere; Google says rankings are not directly affected.\n- Validate all markup, especially anything generated by an LLM, and keep it consistent with visible page content.\n\n## FAQ\n\n**Which schema types should I add first for LLM visibility?**\nS1 suggests Article, FAQPage, HowTo, Organization, Person, DefinedTerm and BreadcrumbList, mapped to templates. S4 highlights FAQPage, Article, Organization, Person and WebPage. Start with Article and Organization, then add FAQPage where answers are self-contained.\n\n**Do LLMs actually read JSON-LD?**\nMicrosoft Bing's Fabrice Canel confirmed in March 2025 that schema helps its LLMs understand content. S6 found schema on around 65% of AI Mode-cited pages and 71% of ChatGPT-cited pages. Other platforms have no equivalent confirmation in these sources.\n\n**Does JSON-LD improve Google rankings?**\nS8 reports that John Mueller confirmed in 2025 that structured data does not directly influence rankings. It still supports rich results and user engagement, and helps search systems classify content.\n\n**Can I let an LLM generate my JSON-LD?**\nOnly with validation. S5 found 40%–50% of GPT-3.5 and GPT-4 markup was invalid, non-factual or non-compliant. Filtering by specialized LLM agents improved quality, and GPT-4 then outperformed humans. Run validator.schema.org and the Rich Results Test on output.\n", "sources": [ { "title": "Schema for AI: Schema.org & JSON-LD for LLM Citations (2026)", "url": "https://alicelabs.ai/en/insights/schema-org-for-ai" }, { "title": "Schema & JSON-LD for LLM Search: Structured Data for LLMO / AEO / GEO | Szymon Słowik", "url": "https://www.szymonslowik.com/json-ld-for-llm-seo" }, { "title": "Structured Data, Not Tokenization, is the Future of LLMs", "url": "https://www.schemaapp.com/schema-markup/why-structured-data-not-tokenization-is-the-future-of-llms" }, { "title": "How Can Schema Markup Specifically Enhance LLM Visibility | Walker Sands", "url": "https://www.walkersands.com/about/blog/how-can-schema-markup-support-llm-visibility" }, { "title": "LLM4Schema.org: Generating Schema.org Markups With Large Language Models - Minh-Hoang Dang, Thi Hoang Thi Pham, Pascal Molli, Hala Skaf-Molli, Alban Gaignard, 2025", "url": "https://journals.sagepub.com/doi/10.1177/22104968251382172" }, { "title": "Structured Data for SEO and LLMs | Schema.org Markup Guide", "url": "https://seranking.com/blog/structured-data" }, { "title": "Schema & Structured Data for LLM Visibility: What Actually Helps?", "url": "https://www.quoleady.com/schema-structured-data-for-llm-visibility" }, { "title": "Structured data: SEO and GEO optimization for AI in 2026", "url": "https://www.digidop.com/blog/structured-data-secret-weapon-seo" } ], "related": [ { "name": "Generative Engine Optimization (GEO): A Primer", "url": "https://agentflare.org/research/generative-engine-optimization-geo-a-primer.html" }, { "name": "llms.txt: The Standard for AI-Readable Sites", "url": "https://agentflare.org/research/llmstxt-the-standard-for-ai-readable-sites.html" }, { "name": "Pay-Per-Crawl and the New Bot Web", "url": "https://agentflare.org/research/pay-per-crawl-and-the-new-bot-web.html" }, { "name": "Crawler Policy on the Modern Web: robots.txt and Beyond", "url": "https://agentflare.org/research/crawler-policy-on-the-modern-web-robotstxt-and-beyond.html" }, { "name": "HTTP 402 & x402: How AI Agents Pay for Content", "url": "https://agentflare.org/research/http-402-x402-how-ai-agents-pay-for-content.html" }, { "name": "The AI Agent Economy in 2026", "url": "https://agentflare.org/research/the-ai-agent-economy-in-2026.html" } ], "pricing": { "price_usd": 0.02, "method": "402", "endpoint": "https://esa.aisa.one/api/v1/access/verify", "autopay_hint": "set crawlerAutoPrice=true with the X-Crawler-Token header", "onboarding": "https://esa.aisa.one/agent/guide.html" }, "powered_by": "AISA — agent-native search, settlement & delivery (https://aisa.one)" }