Schema markup for the AI era: what still matters

Not all schema is created equal anymore. A modern guide to the structured data that actually moves the needle for both Google and LLM citation.

Every few years someone announces the death of schema markup. They’re wrong every time, and they’re especially wrong now. In the AI era, structured data has become more important, not less — but what matters has shifted significantly.

Schema types that matter most

Article and BlogPosting — Still foundational. Ensure datePublished, dateModified, author, and publisher are all present.

FAQPage — Directly feeds AI Overview extraction. If you have a genuine FAQ section, mark it up. Do not manufacture FAQs for schema purposes.

HowTo — For step-by-step guides, HowTo schema dramatically improves extraction accuracy in AI-generated summaries.

Organization and Person — Entity schema is the most underused and highest-leverage type. A fully built-out Organization schema with sameAs references to LinkedIn, Crunchbase, and industry directories signals entity authority to every AI system that encounters your brand.

The principle

Schema is a machine-readable claim about your content. Make claims that are true, structured correctly, and backed by genuine depth — and they’ll compound.