How to Make Your Site Cited by ChatGPT: A GEO Checklist

GEO (Generative Engine Optimization) is the practice of structuring your site so AI answer engines can extract and cite it, the same way SEO structures a site for search rankings. AI engines cite sources that are entity-clear, structured, factual, and self-contained — here's the checklist we run against our own site and every client's.

1. Structured data (highest-leverage single change)

Add JSON-LD for your Organization, your Service offerings, and an FAQPage for any Q&A content. This is the machine-readable layer AI crawlers extract from directly — without it, an engine has to infer your entity and offerings from prose, which it does far less reliably.

2. An llms.txt file

Publish a plain-text, structured summary of your site at /llms.txt — what you do, key facts, and links to your most important pages. It's an emerging convention specifically for AI crawlers, and it costs an afternoon to write.

3. Answer the question in the first two sentences

AI engines extract passages, not whole pages. If your answer to "what is X" or "how do you do Y" is buried three paragraphs into throat-clearing, it won't get pulled into a citation. Open every page and section with a direct, quotable answer, then elaborate below it.

4. Visible FAQ content, not just FAQ schema

FAQ schema that isn't also rendered as visible text on the page is a common mistake — some engines discount or ignore schema-only content because it can be used to game results. Every question in your FAQPage schema should be a real, visible heading with a real, visible answer.

5. Clear entity signals — who is saying this, and why should anyone trust it

Named founders, credentials, and concrete facts (years of experience, companies worked at, projects shipped) give an AI engine — and a human reader — a reason to trust the claim. Vague "our team of experts" language is exactly what AI engines are trained to discount in favor of specific, checkable claims.

6. Comparison and decision content

Content that compares options ("X vs. Y: how to choose") gets cited disproportionately, because it's exactly the kind of question people ask an AI assistant instead of searching for. See our own agency vs. in-house vs. AI tools piece as an example of the format.

7. Don't block the AI crawlers

Check robots.txt explicitly allows GPTBot, PerplexityBot, ClaudeBot, Google-Extended, and CCBot. Some site configurations block these by default (often copied from a template built before GEO was a consideration) — if you want citations, they have to be able to crawl you at all.

8. Keep facts consistent everywhere

An AI engine cross-references claims across your site. If your homepage says "8+ years" and your about page says "10+ years," that inconsistency reduces confidence in every other claim you make. Audit for this the same way you'd audit for a broken link.

The quick version

If you do only three things: add Organization + FAQPage JSON-LD, publish llms.txt, and rewrite your key pages so the first two sentences directly answer the obvious question. Everything else compounds from there.

This is the exact checklist we run against our own site — check our llms.txt and structured data as a live example. If you want this done for your site, tell us what you're building.