AI Search Rankings

How to rank in AI search results

Ranking in AI search means being cited in generated answers rather than placed in a list of links. ChatGPT, Perplexity, Gemini and Google AI Overviews all select sources the same basic way: they need to crawl your pages, extract a direct answer and trust who wrote it. Here is what actually works.

RRonnie Huss

Updated

Video transcript

Your site can rank number one on Google and still be completely invisible to ChatGPT. These are two different systems with two different rulebooks. Ranking is not the same as being cited. We asked six engines one buying question. The answer named somebody else. Of a million websites we scored, two hundred and sixteen were ready for AI search. That is one in four thousand six hundred. And almost all of the rest rank perfectly well on Google. So what actually moves it? It is not backlinks. The heaviest signal in the score is E E A T content, at twenty four percent. Who wrote this. What proves it. Can a machine quote it. Fix those three and you are ahead of nearly everyone.

If you want your website to appear in AI-generated answers, you need to understand a fundamental shift: AI search engines do not rank pages in a list. They synthesise an answer and cite a handful of sources. “Ranking” in AI search means being one of those cited sources.

This guide covers exactly what drives AI search rankings, how it differs from traditional SEO, and the practical steps you can take to improve your visibility across ChatGPT, Perplexity, Gemini and Google AI Overviews.

What “ranking” means in AI search

In traditional Google search, ranking means appearing in the list of ten blue links. Position one gets the most traffic, position ten gets the least. It is a positional hierarchy.

AI search works completely differently. When a user asks ChatGPT, Perplexity or Gemini a question, the AI reads multiple sources, synthesises an answer, and cites two to five websites as references. There is no page two. There is no position seven. You are either cited in the answer or you are not.

This means the goal shifts from “rank higher than competitors” to “be one of the few sources the AI chooses to cite.” The bar is higher in one sense (fewer slots) and lower in another (you do not need to outrank everyone on every keyword, you just need to be the most citable source for specific questions).

Key takeaway: AI search ranking is not about position. It is about being cited, mentioned or quoted inside the generated answer.

The 6 things that drive AI search rankings

Through analysis of thousands of websites and their AI visibility scores, six factors consistently determine whether a site gets cited by AI engines:

1. Readable content

AI engines need to parse your content cleanly. Plain text in HTML, well-structured headings, short paragraphs and clear question-answer formats all help. Content trapped in images, PDFs or interactive JavaScript components is often invisible to AI crawlers.

Structured data can add machine-readable semantics. FAQPage can describe visible FAQs, Article can describe editorial content and Organisation can describe the business entity. Those types do not certify credibility or guarantee citation. Use schema markup where it accurately matches the page.

2. AI crawler access

Your robots.txt file controls which bots can access your site. Many websites - knowingly or not - block the very crawlers that power AI search:

  • OAI-SearchBot (ChatGPT search; GPTBot is OpenAI’s separate training crawler)
  • Claude-SearchBot (Claude / Anthropic search; ClaudeBot is the training crawler)
  • PerplexityBot (Perplexity)
  • Googlebot (Google Search, and the index AI Overviews cites from)
  • Googlebot (Google Search, including eligibility for AI Overviews and AI Mode)
  • Google-Extended is a separate Gemini training/grounding control, not a Search crawler

If a search/discovery crawler the provider relies on is blocked, that can remove a page from that discovery path. Across 6,944 websites audited interactively by SearchScore (July 2026), 6.9% blocked at least one major AI-related crawler or control checked by the audit. Make access decisions by documented role rather than allowing every AI-named token by default.

3. Brand mentions on cited sources

AI engines learn which sources to trust partly from what other trusted sources say. If ChatGPT already cites a particular website in your industry, getting your brand mentioned on that website - through guest posts, expert commentary, interviews or directory listings - increases the likelihood that the AI will encounter and cite your brand.

This is the AI equivalent of link building, but the currency is mentions and co-citations rather than backlinks.

4. Clear authorship and credentials

AI engines favour content from identifiable experts. A page written by “Admin” or “The Team” carries far less weight than one written by a named author with verifiable credentials.

Use Person schema markup on author pages. Include credentials, professional memberships and relevant experience. Link each article to its author page. This gives AI engines confidence that your content comes from a genuine expert, not a content mill.

5. Consistent entity information

AI engines try to understand your brand as an entity - a real, verifiable thing in the world. If your business name, address, services and contact details are consistent across your website, Google Business Profile, industry directories and social media, the AI can confidently identify you.

Inconsistencies - different business names on different platforms, old addresses still listed, services described differently everywhere - create confusion and reduce citation confidence.

6. Content that answers specific questions

AI search engines are question-answering systems. They look for content that directly answers questions users actually ask. Generic content stuffed with keywords does not help. Specific, genuinely useful answers do.

The best approach: identify the real buying questions in your category (use tools like Answer the Public, People Also Ask boxes, or your own customer enquiries), then write clear, complete answers to each one. Each answer should be self-contained, factual and useful on its own.

The difference between Google ranking and AI ranking

SEO and AI search optimisation share a foundation, but the signals that drive each are meaningfully different:

Signal Google SEO AI Search
Backlinks Critical ranking factor Helpful but secondary
Keyword density Important Less relevant
Structured data / schema Useful for supported features and semantics Useful semantic context where consumed
Brand mentions and independent corroboration Authority/context signal Frequently useful evidence source; causal weight varies by engine
Author credentials and EEAT Important for YMYL Critical for all content
AI crawler access (robots.txt) Not applicable Foundational
Content format (Q&A, lists, tables) Helpful Highly valued
Page speed / Core Web Vitals Ranking factor Less critical

The practical implication: a page can rank in position one on Google and still be invisible to AI search engines if it blocks AI crawlers or lacks structured data. Conversely, a well-structured page from a lesser-known brand can get cited by AI engines if it provides the clearest, most authoritative answer to a question.

Step-by-step: how to improve your AI search ranking

Step 1: Run a free audit

Start by understanding where you stand. Run a free AI visibility audit to see your overall score, which AI crawlers you are blocking, what structured data is missing, and how your brand authority compares to competitors.

The audit gives you a prioritised list of fixes so you know exactly what to work on first.

Step 2: Fix AI crawler access

Check yourdomain.com/robots.txt and your edge/WAF rules. If ChatGPT search visibility matters, verify OAI-SearchBot is not accidentally blocked; do the equivalent for Claude-SearchBot and PerplexityBot where relevant. Google Search and its AI features use Googlebot, while Google-Extended is a separate training/grounding control.

Correcting a real search-crawler block removes a hard eligibility barrier. It is high priority when present, but it is not automatically the highest-impact change on a site whose discovery paths are already open.

Step 3: Add structured data

Implement schema markup across your site:

  • Organisation schema on your homepage (business name, logo, contact details, sameAs links to social profiles)
  • Article schema on blog posts and content pages
  • FAQPage schema on pages that answer questions
  • Person schema on author pages, with credentials and biographical details

Learn more about technical GEO and structured data.

Step 4: Write content that answers real questions

Audit your key pages. Do they answer specific questions clearly? If a user asked ChatGPT “what is [your topic]” or “how do I [your task]”, would your page provide the best possible answer?

Focus on:

  • Clear, direct answers in the first paragraph
  • Supporting evidence and examples
  • Original data or insights where possible
  • Named expert authors with credentials
  • Proper formatting (headings, lists, tables) that makes content easy to parse

Read our guide to writing content for AI citation for detailed techniques.

Step 5: Build brand mentions on sites AI already cites

Find out which websites AI engines already cite in your industry. Search ChatGPT and Perplexity for questions related to your niche and note which sources appear most frequently.

Then get your brand mentioned on those sites:

  • Pitch expert commentary to industry publications
  • Contribute guest articles
  • Get listed in relevant directories and “best of” lists
  • Participate in industry roundups and interviews
  • Ensure your Wikipedia or Wikidata entries are accurate where applicable

Each mention increases the chance that AI engines encounter your brand as a trusted reference.

Step 6: Track your results over time

AI search visibility is not a one-and-done fix. It changes as AI models update, as new content is published, and as competitors begin optimising for AI search.

Use SearchScore Tracker to monitor your AI citations weekly across ChatGPT, Claude, Gemini, Perplexity, Grok and DeepSeek. You will see when you start appearing, which queries cite you, and how your visibility trends over time.

Common mistakes that stop you ranking in AI search

Blocking AI crawlers in robots.txt. This is the most damaging issue when present. Across 6,944 websites audited interactively by SearchScore (July 2026), 6.9% block at least one major AI crawler. If the crawler cannot read your site, nothing else matters.

Content only in images or PDFs. AI crawlers read HTML text. If your key information exists only as an infographic, a scanned PDF or a JavaScript-rendered widget, it is effectively invisible. Always provide the same information as plain, structured HTML text.

Missing or inaccurate structured data. Organisation, Article, FAQPage, Person and other valid types can give compatible parsers explicit machine-readable context. Add them when they match visible content, but do not assume that schema alone significantly increases citation likelihood.

Generic content that does not answer specific questions. AI search engines are question-answering systems. If your content reads like a brochure rather than a helpful answer, it will not get cited. Write for real questions, not for keyword density.

Inconsistent brand information. If your business name, address and service descriptions vary across platforms, AI engines cannot confidently identify you as a single entity. Audit your listings and standardise everything.

How long does it take to rank in AI search?

AI visibility has no single propagation speed. Technical changes can matter once the relevant crawler or upstream search index revisits the page. Structured data and llms.txt do not create a separate guaranteed fast lane, and Google indexing itself can range from fast to slow depending on the site and page.

Content and authority changes often take longer because they depend on recrawling, retrieval and external sources changing. Do not promise that a new article will enter AI citations within a particular number of weeks; measure the target questions and record when movement actually occurs.

The six steps above improve readiness and remove avoidable barriers, but SearchScore does not have a controlled basis for promising a four-to-eight-week citation improvement. Track the same questions over time so the outcome is measured rather than assumed.

Compare this to traditional SEO, where a new page can take three to six months to rank - if it ranks at all. AI search rewards speed of implementation.

Check your AI visibility

Free audit. Instant results. No sign-up required.

Check my score

Related articles

Sources & Further Reading

See whether your fixes worked

The Tracker re-checks your site every week across ChatGPT, Claude, Gemini, Perplexity, Grok and DeepSeek, and shows which scores moved. Start with the free scan, then open a Founders Tracker account from your fix list. No card.

Run my free scan →

Frequently asked questions

How do AI search engines decide what to cite?

Answer engines use different retrieval and selection systems, so there is no universal published list of ranking factors. Practical work starts with eligibility and evidence: make pages crawlable and indexable, answer the question clearly, keep entity information consistent, and build independent corroboration. Structured data can add semantic clarity but should not be treated as a guaranteed citation signal.

Is ranking in AI search different from SEO?

Yes. Traditional SEO optimises for position in a list of links, driven heavily by backlinks, keyword relevance and technical performance. AI search optimisation focuses on being cited inside a generated answer, driven by structured data, author authority, brand mentions and question-answering content quality. The two disciplines overlap but require different strategies and different success metrics.

How long does it take to appear in AI answers?

There is no dependable four-to-eight-week timetable. A technical change can matter after the relevant crawler or upstream search index refreshes the page, while authority work can take longer. Measure the same buyer questions repeatedly and report observed movement rather than promising a fixed response window.

Do I need to publish new content for AI search?

Not necessarily. Your existing content may already be high quality - it just needs to be made readable for AI engines. Adding structured data, unblocking crawlers and reformatting key pages into clear Q&A structures can transform existing content into citable material. However, identifying questions your audience asks that you do not currently answer, and creating targeted content for those, will improve your AI visibility further.