How to optimize your site for ChatGPT and Perplexity
Customers no longer just search Google - they ask ChatGPT and Perplexity directly. The new question is not "what position do we rank" but: does AI mention you when it answers? Here is the content checklist, and the technical layer most sites silently fail.
To get cited by ChatGPT, Perplexity and AI Overviews (GEO - Generative Engine Optimization), content needs: a direct answer right at the top, quotable structure (clear headings, short paragraphs, lists, FAQ), schema markup, credible cited sources, and brand signals across many places. Then the technical layer: robots.txt must not block GPTBot, OAI-SearchBot and PerplexityBot; the site must be indexed by Bing (ChatGPT search leans on it); and an llms.txt file gives AI a curated map of your content.
How does AI choose which sources to cite?
- Clear answers: passages that answer the question directly, quotable verbatim.
- Good structure: clear headings, short paragraphs, lists, an FAQ section - easy for AI to parse into ideas.
- Structured data (schema): mark up articles, FAQs, organization and authors so machines understand what is what.
- Credible citations: numbers and claims with named sources raise trust.
- Brand signals: consistent mentions across many places help AI "remember" the brand.
This differs from classic SEO, which optimizes a link's position on a results page; GEO optimizes for being quoted inside the answer. The two complement each other but are not the same - see the full comparison in GEO vs SEO and the foundation in what GEO is.
What is the content checklist?
- Open every article with a passage that answers the main question directly.
- Break content into clear headings, short paragraphs, lists and an FAQ section - the craft of quotable Q&A pairs is in how to write FAQs AI will cite.
- Add schema (BlogPosting, FAQPage, Organization) to every page - details in schema markup for AI.
- Cite clear sources for important numbers and claims.
- Publish an llms.txt to guide AI crawlers.
- Track whether your brand gets mentioned - the method is in checking your brand on ChatGPT.
The technical layer: open the door to the right bots
Great content is worthless if AI platforms' crawlers cannot reach your website. Three checks, 15 minutes:
ChatGPT uses GPTBot (training) and OAI-SearchBot (real-time search); Perplexity uses PerplexityBot. Many sites block all three via a firewall or default configuration without the owner ever knowing - check your robots.txt file and access logs.
ChatGPT's search feature leans partly on Bing's index, so register with Bing Webmaster Tools and enable IndexNow so new content gets picked up fast. This is the most-skipped step, because everyone only tends Google.
A plain-text file at your site root summarizing your content and pointing AI to your key pages - a robots.txt written for language models. An emerging convention, cheap to add, and this site uses one.
Run all of these checks in one pass with our free SEO + GEO Audit tool - it scores any URL on 27 criteria including AI-crawler access, schema and llms.txt, the same bar this site holds itself to at 100/100.
Where do you start?
Write to answer, not to stuff keywords. Each page should clearly answer one real customer question, in a machine-readable structure, with schema and cited sources. That is the foundation on which AI learns to trust and quote you - and because the same foundation lifts Google rankings too, none of this work is wasted even where AI adoption is still ramping.
Checklist from Chạm AI's GEO practice, 2024-2026, including this site's own implementation. Crawler names and platform behavior change; re-verify bot lists quarterly against OpenAI's and Perplexity's official documentation.
Frequently asked questions
How does AI choose which sources to cite?
Five signals matter most: clear answers - passages that answer the question directly and can be quoted verbatim; good structure - clear headings, short paragraphs, lists and an FAQ section that machines can parse; structured data (schema) marking up articles, FAQs, organization and authors; credible citations - numbers and claims with named sources; and brand signals - being mentioned consistently across many places so AI "remembers" the brand.
Which crawlers do ChatGPT and Perplexity use?
ChatGPT uses GPTBot for training data and OAI-SearchBot for real-time search; Perplexity uses PerplexityBot. Many websites block all three through a firewall or default configuration without the owner knowing - check your robots.txt and access logs. ChatGPT's search also leans partly on Bing's index, so register with Bing Webmaster Tools and enable IndexNow.
What is llms.txt and do I need one?
llms.txt is a plain-text file at your site root that summarizes your content and points AI crawlers to your most important pages - like a robots.txt written for language models. It is an emerging convention, not a guarantee of citation, but it costs an hour to create and gives AI tools a clean, curated map of what your site offers. This site uses one.