Every content page on your site should ship Article schema. The minimum useful markup includes headline, datePublished, dateModified, author (with sameAs), publisher (Organization), and image. The author.sameAs field is one of the strongest E-E-A-T signals you can give an LLM.
FAQPage markup must mirror visible content exactly. Google has been flagging mismatched FAQ schema as spam since 2023, and LLMs simply skip pages where the markup does not match what a human would see. A clean visible FAQ at the bottom of every guide, marked up with FAQPage, is one of the highest-converting GEO moves we deploy.
On your homepage, ship Organization JSON-LD with name, url, logo, sameAs (LinkedIn, Wikipedia if applicable, Crunchbase, GitHub, Twitter, Facebook), founder, foundingDate, and contactPoint. This is the foundational entity record the LLM uses to disambiguate you from every other organization with a similar name.
No. The right pattern: Organization sitewide, Article on every content page, FAQPage where you have a real FAQ, HowTo where you have real steps, Product/Service on commercial pages. Do not over-mark.
JSON-LD. Both Google and AI engines parse it most reliably. Microdata and RDFa are still valid spec but you get cleaner extraction with JSON-LD.
No single signal causes a citation, but well-structured schema is one of the strongest contributing signals. Pages we add complete Article + FAQPage + Organization markup to typically see 30-60% citation share lift within 30 days.
Several proposed schema extensions (LearningResource, generative content provenance) are emerging in 2026. Adopt only the ones backed by Schema.org core — avoid speculative custom types.