The study
Over 90 days, we tracked 50,000 unique URLs that appeared in responses from ChatGPT, Gemini, and Perplexity across 12,000 commercial intent queries. We then crawled each URL and audited its structured data. The finding was stark: 94.3% of cited pages had at least one valid schema markup type.
Among pages with no structured data at all, the citation rate was less than 0.4%. That gap, 94% vs 0.4%, is the single most actionable data point in AEO today.
Which schema types matter most
Not all schema is equal. Here's what we found ranked by citation frequency:
- FAQPage, present on 71% of cited pages. The single highest-impact schema type. Pages with FAQPage JSON-LD were cited 3.2x more often than equivalent pages without it.
- Article / BlogPosting, present on 68% of cited pages. Helps models classify content type and extract author credentials for E-E-A-T scoring.
- Organization, present on 61% of cited pages. Signals entity authority to LLMs, particularly important for brand-level citations.
- HowTo, present on 44% of cited pages. Disproportionately effective for instructional queries ("how to", "step by step", "guide to").
- Product, present on 38% of cited pages. Critical for e-commerce and SaaS comparison queries.
- BreadcrumbList, present on 52% of cited pages. Improves content hierarchy understanding for retrieval models.
Why schema helps AI models
Large language models with retrieval capabilities (Perplexity, GPT-4o Browse, Gemini) parse structured data as a high-confidence signal for content type, author authority, and answer structure. When a model is deciding between two pages covering the same topic, the one with FAQPage schema provides machine-readable answers that can be extracted directly, the other requires the model to infer the answer from prose.
The inference step introduces uncertainty. Schema removes it. That's why the citation rate difference is so extreme.
FAQPage schema is not just for FAQ pages. You can add FAQ blocks with JSON-LD to product pages, service pages, and blog posts. Each question-answer pair becomes a directly extractable unit for AI retrieval systems.
How to implement FAQPage schema
Add this JSON-LD block to the <head> of any page with question-answer content:
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [{
"@type": "Question",
"name": "What is AEO?",
"acceptedAnswer": {
"@type": "Answer",
"text": "AEO (Answer Engine Optimisation) is the practice of structuring content so AI engines cite your brand in generated answers."
}
}]
}Quick wins, implement in order
- Add FAQPage schema to your top 10 highest-traffic pages this week.
- Add Organization schema to your homepage if it is missing.
- Add Article schema to every blog post with a clear author byline.
- Add HowTo schema to any guide or tutorial content.
- Validate all schema using Google's Rich Results Test before deploying.
Run an NotionCue scan before and after implementing schema. In our internal tests, sites that added FAQPage schema to their top pages saw a measurable citation increase within 14–21 days.