Get Optimised The GEO industry’s memory

Claim File First published 27 July 2026 Last reviewed 27 July 2026

Does schema markup actually help get my site cited by AI search engines?

The answer

Schema markup does not reliably cause an increase in AI citations. Despite years of promotion, controlled tests show little to no lift.

Verdict: no proven lift in citations Confidence: moderate

The claim

Schema markup has been promoted as a foundational tactic for generative engine optimization since at least 2012. The core claim: adding structured data (especially FAQPage, HowTo, Article, and Organization schema) helps large language models understand and cite a page. By 2026, over 143 promotional articles had been published, with only 18 expressing doubt. The narrative became so dominant that many SEO tools now include schema as a default recommendation for AI visibility.

The receipts

The evidence for schema's impact on AI citations is almost entirely correlational. The most cited number comes from a BetterAISearch study in April 2025: 81% of AI-cited pages use schema markup. But correlation is not causation. The first controlled experiment arrived in May 2026, when Ahrefs tracked 1,885 pages that added schema markup and measured their AI citation rate before and after. The result: citations barely moved. That finding was echoed by Search Engine Journal and later by independent tests from Sortino and others. Meanwhile, promotional content continued at a high pace: 138 pieces were published after the first doubt article in June 2024.

1885

Pages tracked in Ahrefs' controlled experiment on schema markup and AI citations.

The Ahrefs study is the closest thing to a controlled trial in the GEO space. It added schema to 1,885 pages and measured AI citation rates before and after. The result: no statistically significant change. This directly contradicts the promotional narrative that schema is a key driver of AI visibility.

By contrast, the 81% correlation figure from BetterAISearch is often cited as proof, but it does not control for other factors. Pages that already invest in SEO are more likely to use schema, so the correlation may simply reflect overall quality.

What this means for you

Founders should treat schema as a hygiene factor, not a growth lever. Correct schema can prevent errors and may help with traditional rich snippets, but it does not appear to drive AI citations on its own. The real drivers are likely content relevance, topical authority, and citation frequency from other sources. Invest time in those instead.

  1. Audit schema for correctness, not for citations. Use Google's Rich Results Test or Schema.org validator to fix errors. Broken schema can hurt traditional rich snippets, but fixing it will not boost AI citations. Treat it as a baseline requirement, not a strategy.
  2. Invest in topical authority and content depth. AI models cite pages that are frequently referenced by other sources and that cover a topic comprehensively. Build pillar content, earn backlinks, and ensure your content answers the questions your audience asks.
  3. Monitor your own citation sources. Use tools like Brandcited or Mention to track which pages get cited by ChatGPT, Perplexity, and Gemini. Compare pages with and without schema to see if schema correlates with citations in your specific niche. If it does not, reallocate effort.

Confidence, and what would change our mind

Confidence is moderate. The promotional volume is enormous, but the controlled evidence is sparse and negative. The Ahrefs test is the strongest signal we have, but it is a single study. Until we see replication or a mechanism by which LLMs actually parse schema differently from plain text, the prudent stance is that schema is not a causal factor.

Sources