Short answer: no AI platform publishes a ranking formula, and there is no secret signal. Visibility in AI search is influenced by the same broad factors that shape search quality – whether a page can be accessed and indexed, how relevant and clear it is for the question, how specific and well-evidenced it is, how trustworthy its source appears and how consistently the brand behind it is described across the web.
A note on “ranking factors”
Lists of “200 ranking factors” have always been partly guesswork. For AI search the uncertainty is greater: answers are generated, vary between runs and draw on several sources at once. The factors below are therefore grouped by what we can reasonably infer from official documentation, from how retrieval-based systems work and from observing which pages are cited in practice. Treat them as priorities, not a formula.
1. Access and eligibility
This is the only factor that is binary. If a crawler cannot fetch and render a page, it cannot be indexed, retrieved or cited.
- Indexing – Google states that pages must be indexed and eligible to show a snippet to appear as supporting links in AI features.
- Crawler access – OAI-SearchBot for ChatGPT search, PerplexityBot for Perplexity, Googlebot and Bingbot for their ecosystems. See Robots.txt for AI Crawlers: Search Bots vs Training Bots.
- Rendering – content must appear in the HTML crawlers read, not only after scripts run.
- Snippet controls –
nosnippetand similar directives also limit AI features.
2. Relevance to the question
AI systems often break a question into sub-questions and look for passages that answer each one. Pages that address the exact question – using the vocabulary people use, covering the natural follow-ups and matching the intent – are more likely to be retrieved. Broad pages that mention a topic without answering anything specific are weaker candidates.
3. Clarity and extractability
Retrieval systems work with passages. A section that opens with a direct answer, stays on one topic and makes sense when read alone is easier to use than a long narrative. Tables, numbered steps and definitions help because they package information in predictable shapes. See How to Write Direct Answers That Search Engines Extract.
4. Specificity and information gain
If your page says what every other page says, an AI system has no particular reason to use yours. Specific numbers, named methods, concrete examples, local detail and honest limitations add information the system cannot get elsewhere. Research on generative engine optimization has found that adding statistics, quotations and citations to credible sources tends to increase visibility in generated answers.
5. Evidence and first-hand experience
Google’s guidance on helpful content emphasises experience, expertise, authoritativeness and trustworthiness. Content that shows first-hand experience – “we tested”, “in our audits we typically find” – and supports claims with sources is safer to rely on. See E-E-A-T for AI Search: Experience, Expertise, Authority, Trust.
6. Source authority
Systems prefer sources that are trusted for a topic. Traditional signals such as quality backlinks still indicate authority. So do mentions in respected publications, citations by other experts and a history of accurate content in the same field. Topical depth – many good pages on related subjects – strengthens the signal. See Topical Authority: How to Build It the Right Way.
7. Entity clarity and consistency
When search systems can confidently identify who publishes a page and what that organisation is known for, attribution becomes easier. Consistent names, descriptions, founders, locations and profiles across the web, connected with Organization and Person schema, reduce ambiguity. See What Is Entity SEO and Why Does It Matter for AI Search?.
8. Freshness
For topics that change – software, regulations, prices, AI itself – current information is more useful. Clear “last updated” dates, refreshed statistics and removal of outdated advice help. Updating a date without changing content does not.
9. Corroboration across the web
AI systems combine sources. When several trusted sites describe your brand or your facts consistently, a system can state them with more confidence. Contradictions – an old price on a directory, a different service list on a profile – weaken the picture.
10. Page experience
Fast, stable, mobile-friendly pages are part of Google’s overall quality picture and matter for the humans who click through from AI answers. Page experience is rarely decisive alone, but poor experience undermines everything else.
Factors that matter less than people think
- Special “AI schema” – there is none. Structured data helps clarity, not placement.
- Keyword density – repeating a phrase does not make a passage more useful.
- Word count alone – long pages help only if the length adds substance.
- llms.txt – a proposed convention with uneven support; harmless but not a ranking lever. See What Is llms.txt and Should You Use It?.
Factors to avoid manipulating
Hidden text aimed at AI systems, fake reviews, invented statistics and mass-produced pages may seem tempting. They break platform policies, erode trust with the humans who read your content and can be detected and discounted. The durable path is to be the most useful, accurate source for the questions you want to own.
How to prioritise the factors
- Fix access first – it caps everything else.
- Improve clarity on money pages – fast and cheap, with visible results.
- Add specificity and evidence – where your expertise is strongest.
- Build entity consistency – a few days of focused clean-up.
- Invest in authority and corroboration – ongoing, compounding work.
How to test what works for you
Because no formula is published, measurement matters. Track a fixed prompt set, change one group of pages at a time and compare citation rates before and after. Over a few months you will learn which factors move visibility in your market. See How to Measure AI Search Visibility.
Frequently asked questions
Are AI ranking factors the same on every platform?
The principles overlap, but each platform retrieves differently – Google AI features use Google’s index, ChatGPT search uses its own retrieval, Perplexity crawls with PerplexityBot.
Do backlinks still matter for AI search?
Yes, as one signal of authority among several, alongside mentions and entity clarity.
What is the single most important factor?
Access – without it nothing else counts. After that, clear and specific answers.
Next steps
Artificial AI SEO audits every factor above in its AI SEO Audit and prioritises the fixes. Learn more about our approach on AI SEO and GEO.
