Home · SEO Suite · Pricing · Questions · Blog · Contact

How to rank in ChatGPT (and Gemini and Perplexity).

ChatGPT does not have a list of ten blue links, so "ranking" means something different: being named in its answer, being the first brand it recommends, and being a source it links. This guide shows, step by step, what you can actually do to get mentioned by ChatGPT and the other AI assistants, and how to check that it works.

Short answer: To rank in ChatGPT, make sure its crawlers can read your site, publish pages that answer your customers' questions clearly in the first lines, rank well in normal search for those questions, and get mentioned on the sites ChatGPT already cites for your topic. Then ask your key questions regularly on ChatGPT, Gemini and Perplexity to see whether you are named.

Updated 2026-11-02 by the Demandscope team.

What does "ranking in ChatGPT" mean?

In ChatGPT, three things can happen to your brand when someone asks a question about your market: it is named in the answer, it is linked as a source, or it is absent. When several brands are named, order matters too: the first recommendation gets the most attention.

So ranking in ChatGPT means three goals: be named for the questions that matter to you, be named early, and have your pages cited as sources. The same is true for Gemini, Perplexity, Microsoft Copilot and Google AI Overviews.

How does ChatGPT decide what to recommend?

OpenAI does not publish a ranking formula. What is known: ChatGPT answers from what its model learned in training and, for many questions, from a live web search it runs at the moment. When it searches, it reads a few pages and builds its answer from them, and it shows those pages as sources.

That means two paths to being named. The long path: be widely and consistently described on the web, so the model learns about you. The short path: be on the pages ChatGPT reads when it searches today, either your own pages or third-party pages that mention you.

How to get mentioned by ChatGPT: step by step

Pick the questions you want to win: Write 15 to 30 questions your buyers ask: "best [product type] for [audience]", "[competitor] alternatives", "[you] vs [competitor]", "how to [job your product does]". Be specific: a small business will not win "best CRM", but can win "best CRM for plumbing companies".
Ask them and write down what you see: Ask each question in ChatGPT (with search on), Gemini and Perplexity. Note which brands are named, in what order, and which pages are cited. These cited pages are your map: they are what the assistant trusts today.
Open your site to AI crawlers: In robots.txt, allow OAI-SearchBot and ChatGPT-User (the bots ChatGPT uses to search and read pages), and PerplexityBot and Google crawlers if you want those assistants too. GPTBot is OpenAI's training crawler: allowing it helps future models know you.
Serve readable HTML: Many AI crawlers do not run JavaScript. Open your page with scripts turned off: if the main text is missing, they probably cannot read it either. Server-rendered text, a clear title and a meta description fix this.
Build one strong page per question: For each important question, have a page that answers it directly in the first two or three sentences, then gives details, comparisons, prices and examples. Use the question as a heading. Add an FAQ for the follow-up questions.
Rank in classic search for the same questions: When ChatGPT searches, it often reads pages that already rank well. Classic SEO (good content, internal links, backlinks, technical health) is still the base of being cited.
Get listed where ChatGPT looks: Go back to the pages ChatGPT cited in step 2. Ask to be added to comparison articles, get honest reviews on review sites, answer questions on forums where your buyers talk, and make sure your directory listings are complete.
Describe yourself the same way everywhere: Use one clear sentence about what you do and who it is for, and repeat it on your homepage, about page, social profiles and listings. Consistent facts help the model describe you correctly.
Add llms.txt and structured data: An llms.txt file at your site root lists your key pages for AI tools. Schema markup (Organization, Product, FAQPage, Article) makes your facts machine-readable. Neither is a magic switch, but both remove guesswork.
Check again every few weeks: Ask the same questions again and compare with your first notes. Count how often you are named and cited, per assistant. Improvements usually show on assistants that search live (like Perplexity) before they show in what a model knows by itself.

How to rank in Gemini, Perplexity and Google AI Overviews

The steps are the same, with small differences. Perplexity searches the web for almost every question and always shows its sources, so improvements to your pages and to third-party mentions often show there first. Gemini and Google AI Overviews lean on Google's own index, so ranking well on Google for the question is especially important. Google-Extended in robots.txt controls whether Google may use your content for its Gemini models; it does not change your normal Google ranking.

ChatGPT: allow OAI-SearchBot and ChatGPT-User; be on the pages it cites.
Perplexity: allow PerplexityBot; clear answers and fresh pages help.
Gemini and AI Overviews: rank well on Google for the question.
All of them: third-party mentions and consistent facts.

Which questions should you track?

Track the questions that come right before a purchase, not only general questions about your topic. A brand that is named for "what is invoicing software" gets attention; a brand named for "best invoicing app for plumbers" gets customers. Mix four kinds of questions so you see the full picture.

Best-of questions: "best [product type] for [audience]", "top [product type] in 2026".
Alternative questions: "[competitor] alternatives", "cheaper alternative to [competitor]".
Comparison questions: "[you] vs [competitor]", "[competitor A] or [competitor B]".
Problem questions: "how do I [job your product does]", "what is the easiest way to [job]".
Trust questions: "is [you] legit", "is [you] worth it", "[you] reviews".

A quick checklist before you start

robots.txt does not block OAI-SearchBot, ChatGPT-User, PerplexityBot or Google crawlers.
Your main pages show their text with JavaScript turned off.
Each key page has a clear title, a meta description and a direct answer in its first lines.
Your homepage says in one sentence what you do and who it is for.
An llms.txt file lists your most important pages.
Your brand name, description and prices are the same on your site and your listings.
You have a list of 15 to 30 questions and a first measurement to compare with.

What does not work

Hidden text or keyword stuffing "for AI": assistants read the page like a person would, and search engines penalize hidden content.
Fake reviews and spam on forums: they are often removed and can damage trust in your brand.
Judging from one answer: answers change between runs, so measure many questions several times.
Paying someone who guarantees a position in ChatGPT: nobody can guarantee it.

How to track your ChatGPT ranking with Demandscope

Doing this by hand works for a few questions, but it gets slow. Demandscope's AI Visibility tool asks up to 30 questions to ChatGPT, Perplexity and Gemini and checks Google AI Overviews in one report. It shows whether you are named, your position, which of your pages are cited, your share of voice against up to 5 competitors, and the questions where a competitor was named instead of you.

Site Audit checks your robots.txt and llms.txt and tells you if AI bots are blocked. On the Advanced plan, the same questions are asked again automatically every 7, 14 or 30 days, so you can see your progress over time.

Questions

Can you pay to rank in ChatGPT?

Not in normal answers today. Being named depends on what the model knows and what it finds when it searches. Be wary of services that promise a guaranteed position.

How long does it take to show up in ChatGPT?

If ChatGPT searches the web for your question, changes to your pages and new third-party mentions can show within days or weeks. What the model knows by itself only changes with new model versions, which takes months.

Does ChatGPT use Google or Bing to search?

OpenAI runs its own search crawler (OAI-SearchBot) and also works with search partners. In practice, pages that rank well in classic search are often the ones ChatGPT reads, so classic SEO still matters.

Which robots.txt rules do I need for ChatGPT?

Do not disallow OAI-SearchBot and ChatGPT-User if you want ChatGPT to find and read your pages. GPTBot is used for training; you can allow or block it separately.

Why does ChatGPT recommend my competitor and not me?

Usually because your competitor is described on more of the pages ChatGPT reads: comparison articles, review sites, forums. Look at the sources it cites for your question and work to be included there.

Is ChatGPT SEO the same as SEO?

It builds on SEO but measures something different: mentions and citations in answers instead of positions in a list. Good SEO is the base; ChatGPT SEO adds clear answers, AI crawler access and third-party mentions.