Technical

Schema for LLMs: what structured data actually changes

Which structured data influences retrieval and attribution, and which is mostly decoration.

Structured data does not guarantee a citation, but it removes ambiguity. For answer engines, ambiguity is expensive: an unclear entity or unverifiable claim is easier to skip than to resolve.

The schema that pays

Prioritise markup that answers identity, price and proof questions.

  • Organization with consistent name, logo, sameAs and contact points.
  • Product and Offer with currency, availability and price validity.
  • FAQPage and HowTo for passages models lift most often.
  • Article with author, reviewer and publication dates.
  • Review and AggregateRating where you can substantiate them.

Consistency beats volume

One canonical description repeated across site, schema, directories and press is worth more than five hundred unvalidated types. Audit for contradictions first: conflicting prices or product names are worse than no markup at all.

Validate and monitor

Test in Google Rich Results and schema validators, watch for drift after deploys, and re-check entity resolution in model answers each quarter. Schema is a contract with machines, and contracts decay when nobody re-reads them.

Back to all articles
No lock-in audit

Find out what AI says about your brand

Run the free AI visibility report: we check a panel of buyer prompts across ChatGPT, Perplexity, Gemini, Copilot and Google AI Overviews, then send you the citation gaps worth fixing first.

See pricing