Which schema types actually matter for AEO?
Five carry most of the weight: Organization (who you are — the anchor of your entity graph), WebSite, Article or BlogPosting with real dates (freshness engines can parse), FAQPage (answers in exactly the shape answer engines extract), and BreadcrumbList (site structure). The honest caveat schema vendors skip: structured data mostly helps machines parse and trust what your content already says — it does not make weak content citable, and Google's own guidance is that markup must describe visible page content, not decorate it.
Get the entity graph right before adding page types: one Organization node with a stable @id, referenced by everything else, rather than duplicate organization declarations scattered per page. Engines resolving "who is this" reward consistency; conflicting nodes on different pages are how a brand ends up as two half-known entities.
Use the type that matches what the page IS. Google's guidelines reserve QAPage for forums with user-submitted answers — a company answering its own question is FAQPage. Mislabeled schema is worse than missing schema, because it signals either carelessness or manipulation to the systems reading it.
Validate against what parsers actually see: JSON-LD with malformed syntax fails silently, and a page can carry six schema blocks of which zero parse. Test with Google's Rich Results tooling or by parsing your own JSON-LD in CI — we audit customer sites by parsing the markup, never by guessing from prose, for exactly this reason.
Sources
Last reviewed 2026-08-10. We revisit these when the underlying research changes.
Related
See where you actually stand
Run a free check across ChatGPT, Perplexity, Gemini, Claude and Google AI Mode.
Check your AI visibility