article topic
sections to work through
last updated
AI platform features change quickly. This article describes what was verified on the date of the last update. Check the provider's current documentation before making an implementation decision.
ChatGPT is not a search engine. It does not rank pages, does not crawl them the way Googlebot does and does not build an index. That means most conventional SEO knowledge is only partly useful here. ChatGPT has two sources of knowledge about your company: training data gathered before the model's cut-off, and - in search mode - data fetched in real time. Optimisation has to address both.
How does ChatGPT learn about your company?
OpenAI's models are trained on enormous datasets from the web. If your company was regularly mentioned in articles, discussions and reviews before the training cut-off, the model "knows" about it. If not, it is absent from that knowledge regardless of how good your website is.
In search mode, ChatGPT fetches web data in real time using a search tool. That means pages indexed today can reach ChatGPT answers relatively quickly. GPTBot, OpenAI's crawler, needs access to your site.
Block 1: technical accessibility fundamentals
- Check robots.txt - make sure GPTBot is not blocked. Open yourdomain.com/robots.txt and look for "User-agent: GPTBot" entries with a Disallow directive. If you block GPTBot, ChatGPT cannot read your site.
- Check the others too: OAI-SearchBot, PerplexityBot, ClaudeBot and anthropic-ai. All should have access - either no Disallow entry or an explicit Allow.
- Make sure key content is rendered server-side. AI crawlers do not execute JavaScript, so if your content loads through JS in a single-page app without server rendering, ChatGPT may see an empty page.
- Add an llms.txt file to the domain root - structured information about the company for AI systems.
- XML sitemap - a current site map helps crawlers find every page.
How to check whether GPTBot can reach your site
Open yourdomain.com/robots.txt. A correct file that grants GPTBot access either has no "User-agent: GPTBot" entry at all (default access) or an explicit "User-agent: GPTBot / Allow: /". The red flag is "User-agent: GPTBot / Disallow: /". Many companies blocked OpenAI in 2023 and 2024 out of concern about training-data scraping. If that entry is still there, you are blocking ChatGPT from fetching current data.
Unblocking GPTBot gives ChatGPT access to fetch content in search mode, not to train models. Model training uses separate crawlers such as CCBot. You can block CCBot and still allow GPTBot.
Block 2: site structure and content
- Define your specialisation in the first 100 words of the home page - ChatGPT has to understand immediately who you are and for whom. "We help B2B technology companies improve their visibility in answers from AI systems" is a sentence ChatGPT can cite.
- Add a definition in an "X is..." format for every key service or product on its dedicated page.
- Create an FAQ page with questions phrased as natural buyer queries - a format ChatGPT cites especially readily.
- Every article and page should have a visible author with a full name and publication date. Anonymous content is less credible to AI.
- Add JSON-LD schemas: Organization for company data, Person for experts, Article for articles, FAQ for questions and answers. Schemas help AI understand the structure of the site.
- Make sure the site loads quickly and correctly on mobile - ChatGPT fetches content much as a browser does.
Block 3: content ChatGPT can cite
Good technical access and good structure are necessary but not sufficient. ChatGPT cites content that is informative, specific and well structured. A checklist for every article or service page:
- Does the article answer one specific question a buyer might ask?
- Does the first section give a direct answer within 50 to 60 words?
- Is each paragraph a self-contained thought of 134 to 167 words that can be lifted without losing sense?
- Does the article contain at least one specific fact with an attributed source?
- Are the headings phrased as questions or direct statements?
- Does the content use lists or tables where that is natural?
- Does the article carry a publication date and an update date?
Block 4: external mentions - the most important element
What you do on your own site has a limited effect on what ChatGPT knows from training data. The real breakthrough comes when other sites start talking about your company - industry articles, reviews, discussions and LinkedIn posts. The model learns your company from the context in which others mention it.
| Action | Effect on ChatGPT | Difficulty | Time to effect |
|---|---|---|---|
| Unblocking GPTBot in robots.txt | Medium (real-time search) | Low | Immediate |
| Site content optimisation | Medium | Medium | 2-6 weeks |
| llms.txt | Low to medium | Low | Immediate |
| Articles in industry publications | High | Medium | 4-8 weeks |
| Expert activity on LinkedIn | High | Medium | 4-12 weeks |
| Mentions on Reddit and forums | Very high | Medium | Variable |
| Proprietary data and research | Very high | High | 3-6 months |
The checklist: 15 things to verify now
- 1robots.txt - GPTBot is not blocked
- 2robots.txt - PerplexityBot and ClaudeBot are not blocked
- 3Key content is rendered server-side, not only through JavaScript
- 4llms.txt is present at the domain root
- 5The XML sitemap is current
- 6Home page states a clear specialisation within the first 100 words
- 7Every article has a named author
- 8Every article has a publication date and an update date
- 9JSON-LD Organization schema on the home page
- 10JSON-LD Article schema on article pages
- 11At least five expert articles answering buyer questions
- 12An FAQ on key service pages
- 13At least three mentions on external industry publications
- 14An active LinkedIn profile with expert articles
- 15Google Business Profile reviews describing your specialisation
Assess whether your category is suitable for AI Visibility measurement.
We assess category fit free of charge. Full visibility and competitor measurement is delivered within a Snapshot or a strategic audit.
0 PLN · Fit assessment, without a promise of a full report
Frequently asked questions
How do I check whether GPTBot can reach my site?
Open yourdomain.com/robots.txt and make sure there is no "Disallow: /" entry for User-agent: GPTBot.
Does optimising for ChatGPT differ from SEO?
Yes. SEO focuses on keywords, whereas GEO emphasises structured facts and citable passages.