What "ranking in ChatGPT" actually means
AI search engines do not have a ranked list of results the way Google does. When ChatGPT or Perplexity answers a question, it either retrieves and cites specific pages (retrieval-augmented generation) or draws on training data. Being "ranked" in AI search means being the page that gets cited - appearing as a source in the answer with a link the user can follow.
The criteria for citation are different from traditional SEO ranking: AI engines weight machine-readability, structured data quality, E-E-A-T signals, and unambiguous definitions over backlink volume or keyword density. A well-structured page on a small domain can outrank a major publisher if it answers a query more directly.
Check your current AI search readiness
Before implementing any changes, run a free AI visibility check on your site. It scores 32 signals across AI crawler access, structured data, E-E-A-T, and content structure - and returns a prioritised fix list specific to your domain.
Check AI visibility for free7 signals that determine AI search citation
Allow AI crawlers in your robots.txt
ChatGPT uses GPTBot and OAI-SearchBot. Perplexity uses PerplexityBot. Claude uses ClaudeBot and anthropic-ai. If any of these are blocked - intentionally or by a wildcard Disallow rule left over from an aggressive SEO setup - the engine cannot read your content and cannot cite you.
How to implement this
Open your robots.txt (/robots.txt) and check for Disallow: / or user-agent: * blocks that catch these bots. Each AI search bot needs an explicit Allow: / entry above any broad disallow rules. There are 8 named AI bots worth checking: GPTBot, OAI-SearchBot, PerplexityBot, ClaudeBot, anthropic-ai, CCBot, Bytespider, and Amazonbot.
Add FAQPage structured data to your key pages
ChatGPT and Perplexity both parse JSON-LD schema to understand page content without rendering it. FAQPage schema is the highest-signal structured data type for AI engines because it packages a question and a definitive answer in a machine-readable pair - exactly the format AI answer engines are trying to produce.
How to implement this
Add a FAQPage JSON-LD block to your most important pages. Each Question/acceptedAnswer pair should be a real question your audience types, with a direct answer in the first sentence. Three to seven pairs per page is enough. Avoid questions that begin with "How does [Your Brand] work" - write questions the way a user would type them into ChatGPT.
Write with definitional sentences above the fold
AI engines cite pages that clearly define the topic they cover. A page titled "What is SEO" that spends the first paragraph describing the company before defining SEO will rarely be cited over a page that opens: "SEO (Search Engine Optimisation) is the practice of improving a website's position in organic search results by making it easier for search engines to crawl, understand, and rank its content."
How to implement this
Add one clear "X is a Y that does Z" sentence near the top of any page you want cited. The definition should be complete in one sentence and not require the reader to scroll for context. AI engines extract these sentences directly for use in answers.
Build E-E-A-T signals into your content
E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) was introduced by Google to evaluate content quality, but the same signals determine whether AI engines treat a source as reliable enough to cite. A page with no author attribution, no external citations, and no organisational context scores lower for AI citability than one with all three.
How to implement this
Add a visible author byline (name + brief credential). Link to at least two relevant external authoritative sources. Add an About page that describes who runs the site and their relevant background. Add a Contact page. These are also the signals the AI Search Checker checks under E-E-A-T - a page scoring below 50 on E-E-A-T is rarely cited regardless of content quality.
Structure content with clear H2/H3 hierarchy
AI engines read raw HTML, not rendered pages. A JavaScript-heavy site where all content is loaded client-side may appear empty to a bot reading the raw response. Even for sites that render server-side, content without heading hierarchy is harder to parse: AI systems use H2/H3 tags to understand section scope and identify which part of a page answers a given query.
How to implement this
Use one H1 per page. Use H2 for major sections. Use H3 for sub-points within sections. Avoid using bold text as a substitute for headings. Confirm your H1 is present in the raw HTML (view-source, not DevTools) - if it is missing in raw HTML but visible in the browser, your site uses JavaScript rendering and AI bots may be seeing an empty page.
Create an llms.txt file
llms.txt is an emerging standard (proposed in September 2024) that gives AI systems a plain-text overview of your site - what it does, its key pages, and any usage policy. Perplexity and some other AI tools have confirmed they read it when present. It is not confirmed by ChatGPT or Google's AI systems, but it is a low-effort addition that signals you are AI-aware and provides a clean summary for systems that do read it.
How to implement this
Create a file at /llms.txt. Include: a one-paragraph description of what your site does, a list of your 5-10 most important pages with brief descriptions, and an optional usage policy (whether AI training use is permitted). Keep it under 500 words. This is a 15-minute task that scores 3 points on the AI Search Checker.
Run a full AI search readiness audit
The six signals above are the highest-impact ones, but AI search readiness covers 32 distinct checks across five categories: AI Crawler Access, Structured Data, Content Structure, E-E-A-T Signals, and Technical Signals. Checking them manually takes hours. An automated audit covers all 32 in 20 seconds and returns a prioritised fix list specific to your site.
How to implement this
Run the free AI Search Visibility Checker on your site URL. It checks all 8 named AI bots, FAQPage and HowTo schema, llms.txt, ai.txt, H1/H2/H3 hierarchy, author attribution, About/Contact pages, HTTPS, Open Graph tags, canonical URL, and JavaScript rendering detection - all without requiring a login.
Priority order: where to start
Not all seven signals are equal. AI crawler access is the prerequisite - a blocked bot cannot cite you regardless of how good your structured data is. Fix that first. Structured data (FAQPage schema) has the highest return on time invested after access is confirmed. E-E-A-T and content structure improvements compound over time.
How AI search differs from traditional SEO
Traditional SEO optimises for a ranked list of blue links. AI search produces a synthesised answer, with cited sources that the user can verify. The differences in ranking criteria follow from this:
| Signal | Traditional SEO | AI Search (AEO/GEO) |
|---|---|---|
| Backlinks | Critical ranking signal | Moderate - brand authority matters, volume less so |
| Structured data | Helpful for rich results | High priority - FAQPage pairs are citation units |
| Keyword density | Medium - avoid over-optimisation | Low - clear definitions matter more than repetition |
| Page speed | Ranking factor | Less direct - but slow pages get crawled less |
| Bot access | Googlebot in robots.txt | 8+ named AI bots, each needs explicit Allow |
| Author attribution | Minor E-E-A-T signal | High - AI engines weight authorial credibility |
| Definitions above fold | Not a ranking factor | High - AI engines extract these for direct answers |
Traditional SEO and AI search optimisation are complementary, not competing. Use the free SEO Optimiser to check traditional signals alongside the AI Search Checker for AI-specific readiness.
Frequently asked questions
Does traditional SEO help you rank in ChatGPT?
Partially. Technical SEO signals like HTTPS, canonical URLs, and crawlability matter to AI engines too. But AI search adds new requirements: structured data (FAQPage, HowTo), E-E-A-T attribution, and AI bot access in robots.txt. A site with strong traditional SEO but blocked AI crawlers will not be cited by ChatGPT regardless of its Google rankings.
How long does it take to see results after making AI search optimisations?
AI engines like Perplexity re-crawl regularly. Structural changes (robots.txt, schema, llms.txt) can take effect within days of a crawl. E-E-A-T signals build over time - author attribution and external citations are slow signals. Most users see measurable improvement in citation frequency within 4-8 weeks of implementing the full set of changes.
What is the difference between AEO and GEO?
AEO (Answer Engine Optimisation) is the practice of structuring content so AI-powered answer engines can extract and cite it directly. GEO (Generative Engine Optimisation) is a broader discipline focused on being surfaced in generative AI responses across search, chatbots, and AI assistants. AEO focuses on structured answers (schema, FAQs, definitions). GEO also includes brand authority, citation networks, and content freshness.
Do I need to submit my site to ChatGPT or Perplexity?
No. ChatGPT (GPTBot and OAI-SearchBot) and Perplexity (PerplexityBot) crawl the web independently, similar to Googlebot. You do not submit URLs to them. The only submission-like action is ensuring your robots.txt does not block them, and optionally adding an llms.txt to help them understand your site's purpose.
What is the most common reason a site is not cited by AI engines?
In our analysis of 500+ sites scanned through the AI Search Checker, the most common issue is robots.txt blocking AI crawlers - either via a wildcard Disallow: / rule or by explicitly blocking GPTBot or PerplexityBot. The second most common issue is missing structured data (no FAQPage or HowTo schema). Together these two categories account for over 70% of low scores.
Run your AI search readiness check
The free AI Search Visibility Checker scores your site against 32 signals across AI crawler access, structured data, E-E-A-T, and content structure. Results in 20 seconds. No login required. You get a prioritised fix list with the specific changes most likely to improve your citation frequency in ChatGPT, Perplexity, and Gemini.
Check my AI search visibility