SEO vs GEO: what changed, and what to do about it

Generative engines do not rank pages the way Google does. Most combine a retrieval step, pulling in pages that match the query, with the model’s own judgement about which of those pages to trust and quote. A few patterns show up across ChatGPT, Perplexity and Gemini in practice:

• Clear entities. Pages that name the product, company, person or place plainly, rather than relying on pronouns and vague references, are easier for a model to attribute a claim to.
• Quotable claims with numbers. A sentence with a specific number, date or statistic is easier to lift into an answer than a paragraph of general description.
• Structured data. Schema markup (FAQPage, Article, Organization) gives a model a machine-readable shortcut to the same facts it would otherwise have to infer from prose.
• Author credentials. A named author with a bio and relevant expertise is a trust signal, the same way it is for Google’s own guidelines.
• FAQ-style sections. Question-and-answer formatting maps almost directly onto how a user phrases a prompt, which makes the answer easy to extract.
• llms.txt and robots.txt access. If an AI crawler is blocked in robots.txt, or your site offers no llms.txt pointing to your key pages, some engines simply cannot see you as a candidate source at all.

The SEO Agent covers the Google side: keyword research, content writing, technical audits, schema and internal linking, aimed at ranking in the ten blue links. The GEO Agent covers the generative side: it scores how citable your pages already are, generates the fix files (llms.txt, schema, FAQ sections, author bios) from the checklist above, and tracks whether you actually get cited across six AI engines over time. You can also run a one-off check with the free AI visibility check before deciding whether GEO work is worth prioritizing for your site.