A practical GEO checklist: how to get cited by Google AI Overviews, ChatGPT and Perplexity. Schema, llms.txt, content structure and freshness signals.
AI search engines — Google AI Overviews, ChatGPT, Perplexity and Gemini — increasingly answer questions directly, citing specific pages. GEO (Generative Engine Optimization) is the practice of structuring content so that your page gets picked as the source.
Here is the checklist we apply at WebWind on every project.
Classic SEO has one goal: rank high. GEO has a different one: get cited in an AI-generated answer. A well-structured page can become a source even from page two of Google — models pick passages by quality and structure, not just rank.
| Aspect | SEO | GEO | | --- | --- | --- | | Goal | Rank in the top 10 | Be the cited source | | Format | Any | Direct, structured | | Length | Often long | Concise + complete | | Trust signal | Links | Citations and data |
Models extract single paragraphs, not whole pages. Each heading should be a user question, and the first paragraph under it a self-contained answer (40–60 words).
JSON-LD helps AI understand content type. Minimum: Organization, WebSite,
BreadcrumbList. For content: Article/BlogPosting, FAQPage, and HowTo
for step-by-step guides.
llms.txt — a site map for models (llmstxt.org)./pricing.md, /services.md — machine-readable data an AI agent reads without
rendering JavaScript.Accept: text/markdown.Publish and update dates should be visible, with dateModified in schema. Pages
updated regularly get picked more often than static ones.
[source]).Check robots.txt — GPTBot, PerplexityBot, ClaudeBot, Google-Extended
should have access if you want to be cited.
No. GEO builds on good SEO — the same quality content serves both. The difference is in structure and authority signals.
Run your key questions through ChatGPT, Perplexity and Google, and see whether your domain appears as a source. Repeat monthly.
With structured data and llms.txt — the fastest high-impact wins.
.