AI Search Services Insights Get Answerability

// Structured Data

Schema Markup Still Works. Just Not For This.

What Google's own documentation says about structured data and AI citation, what the best controlled test found, and where schema still earns something real.

Talk to us about your brand’s footprint

The Claim

Schema.org Is Sold as an AI-Citation Lever. It Isn’t One.

Add structured data, get cited more by AI Overviews and ChatGPT, that's the pitch behind a lot of "AI-ready schema" audits being sold right now. It's an easy story: markup helps machines understand your page, AI systems are machines, therefore markup helps AI systems. The actual evidence doesn’t support the conclusion.

What The Evidence Shows

What Google’s Documentation and Independent Testing Actually Say

Google's own Search Central documentation states plainly that no special schema.org structured data is needed to appear in AI Overviews or AI Mode. Standard indexability and a match between your structured data and your visible text are what the documentation actually asks for.

The strongest controlled test available backs this up. Ahrefs tracked 1,885 pages that added JSON-LD schema against 4,000 matched control pages over a 30-day window. The result: no meaningful citation lift on Google AI Mode or ChatGPT, changes within statistical noise, and a small negative movement on AI Overviews that isn’t confidently attributable to the schema change itself. A separate correlational analysis found AI-cited pages were roughly three times more likely to carry JSON-LD, but the controlled test found that relationship isn’t causal, schema tends to co-occur with authority, content quality, and links, not cause citation on its own.

Where It Still Helps

Where Schema Markup Still Earns Something Real

Schema.org markup still does genuine work, just not the work it's currently being sold to do. It remains a real driver of rich results and shopping feed eligibility in classic search. Microsoft has stated that schema helps its own systems (including Bing Copilot) understand content, though independent testing hasn't yet shown that translating into a measurable citation increase.

One further nuance worth knowing: a 2026 paper found that isolated JSON-LD script tags give only marginal accuracy improvement in AI retrieval testing, but a fully integrated approach, structured metadata combined with matching semantic hierarchy and visible content, showed a real accuracy lift. Worth flagging directly: that paper's authors work for a company that sells exactly that kind of structured-data implementation, so treat it as one credible data point, not independent confirmation of the magnitude.

Questions

Frequently Asked

Does adding schema hurt anything?

No documented harm, provided it's accurate and matches your visible content. Structured data that contradicts what's actually on the page is a separate, real risk Google does warn about. Accurate schema doesn’t hurt AI visibility, it just doesn’t reliably help it either.

Should we stop using schema markup?

No. Keep it for what it actually earns: rich results, shopping feed eligibility, and Bing's stated comprehension benefit. Just don’t budget for it as an AI-citation fix.

What’s the difference between "isolated" schema and a structured entity approach?

Isolated JSON-LD without matching visible content showed only marginal improvement in one study. A fully integrated approach, structure plus visible content saying the same thing, showed a larger lift, in a paper written by a vendor that sells that exact implementation. Real finding, but weigh the magnitude accordingly.

When would schema actually be the priority?

If your product feeds or rich-result eligibility are the actual gap, an e-commerce catalogue missing Product schema, a local business missing LocalBusiness markup, fix that first, for the real, confirmed benefit it has in classic search. That’s a different project from AI citation work.

See How Entity and Brand Optimisation Brings This Together

Structured data is one piece of a larger picture, alongside entity recognition and cross-web consistency. See the full approach on the Entity and Brand Optimisation page.

Back to Entity and Brand Optimisation