How to Rank in ChatGPT

To rank in ChatGPT, structure each key page around a clear 40 to 60 word direct answer, add FAQ and Organization schema, publish an llms.txt file, earn citations on trusted industry sources ChatGPT retrieves from, and build topical depth around a defined category. ChatGPT does not rank pages like Google; it summarises and cites based on clarity, structure and source authority.

Why ChatGPT is different from Google

Google ranks pages and shows you a list. ChatGPT reads many pages, synthesises an answer, and names two or three businesses. There is no page 2. If you are not in the synthesis, you are nowhere.

This changes the optimisation game. You are not fighting for position 1; you are fighting to be one of the sources the model considers worth citing.

7 steps to rank in ChatGPT

  • Step 1 - Direct-answer paragraphs: open every commercial page with a 40 to 60 word factual summary. ChatGPT lifts these almost verbatim.
  • Step 2 - FAQ schema: add FAQPage JSON-LD to every key page. Questions and answers are the format LLMs were trained to extract.
  • Step 3 - Organization and Person schema: tell ChatGPT who you are, where you operate, what you do and who runs the business.
  • Step 4 - llms.txt: publish an llms.txt file at the root of your site listing your key pages and what each one covers. ChatGPT, Claude and Perplexity all respect it.
  • Step 5 - Topical depth: publish 6 to 12 deep articles around your category. ChatGPT favours sites it recognises as topical authorities, not one-page businesses.
  • Step 6 - Earn external citations: get mentioned on industry publications, podcasts, Wikipedia, LinkedIn and trusted directories. ChatGPT weights these heavily when deciding who to name.
  • Step 7 - Consistent NAP and entity signals: name, address, phone, company number, founder name. Same data everywhere. Inconsistency makes ChatGPT hedge or skip you.

What does not work

  • Keyword stuffing. ChatGPT does not count keywords; it understands meaning.
  • Thin AI-generated content. Models can detect their own output and weight it down when citing.
  • Hidden text or schema mismatches. Schema must reflect what is actually on the page or it gets discounted.
  • Buying low-quality backlinks. ChatGPT weights citation quality, not quantity.
  • Ignoring off-site presence. Your website alone is not enough. ChatGPT cross-references many sources before naming a business.

How long it takes

ChatGPT re-trains and retrieves on a faster cycle than Google updates rankings. Most businesses we work with start appearing in some prompt-set answers within 4 to 8 weeks of a focused effort. Reaching consistent share-of-voice in a competitive category takes 3 to 6 months.

The single fastest lever is usually earning two or three citations on trusted UK industry publications. That alone often moves a business from zero mentions to regular mentions within a month.

A worked example

We tested the prompt 'best UK AI consultancy for motor claims' across ChatGPT (GPT-4o and GPT-5), Perplexity and Gemini. Three businesses were named in 9 of 12 runs; another six were named occasionally; the rest were never named. The three winners all had: direct-answer paragraphs on their service pages, FAQ schema sitewide, an llms.txt file, and at least three citations on UK insurance publications. The losers had strong Google rankings but none of the above. Same expertise, very different AI visibility.

Frequently Asked Questions

Do I need to pay OpenAI to appear in ChatGPT?

No. ChatGPT does not sell placement in answers. Visibility is earned through content structure, schema, citations and topical authority, exactly as with organic search.

Does ChatGPT use my website in real time?

Sometimes. ChatGPT uses both trained knowledge and live web retrieval (via Bing). Live retrieval reads your current site; trained knowledge reflects what existed at the model's training cut-off. Optimising for both means current pages must be clear, and external citations must exist on sites the model was trained on.

What is llms.txt and where do I put it?

llms.txt is a plain-text file at the root of your domain (yourdomain.com/llms.txt) listing your most important pages and what each covers, in markdown. It helps AI engines understand your site quickly. Anthropic, Perplexity and a growing number of crawlers respect it.

Can I see whether I rank in ChatGPT?

Yes. Run a fixed prompt set monthly across ChatGPT, Perplexity and Gemini and record which businesses are named and which sources are cited. This is the core of an AI visibility audit.

Does blocking ChatGPT in robots.txt help or hurt?

Hurt. If you block GPTBot, you stop ChatGPT from reading your site, which means it can never cite you. Most UK businesses should allow GPTBot and use llms.txt to guide it.

Loading the full page. If it doesn't load, JavaScript is required.