How to Get Cited in ChatGPT and Perplexity - How AI Engines Choose Sources and What to Do

ChatGPT and Perplexity answer questions with ready-made text and list the sources they used underneath. According to OpenAI, more than 900 million people now use ChatGPT every week. For a business, that's a new visibility channel where you don't compete for a position in a list but for being cited. Both engines cite pages they can retrieve, from which a specific answer is easy to extract, and which they trust - and you have real influence over all three.

How ChatGPT and Perplexity choose sources, how they differ (Wikipedia versus Reddit), what research says increases your chances of being cited, step-by-step site preparation and a test kit that checks your visibility in an hour.

In this guide I show how ChatGPT and Perplexity choose sources, how they differ, what research says increases your chances of being cited, and how to check where you stand in an hour with a simple test kit. If you're new to the topic, start with the guide GEO - what it is and how it differs from SEO.

Why it's worth the effort

Traffic from AI assistants is still small in volume, but it's valuable and growing:

• Scale. In February 2026, OpenAI said more than 900 million people use ChatGPT every week.
• Traffic quality. Ahrefs showed with its own 2025 data that visits from AI search made up 0.5% of traffic but produced 12.1% of signups, converting about 23 times better than search engine traffic. That's one company's result, but it shows well that someone recommended by an assistant arrives already convinced.
• The trend. In 2024, Gartner predicted that by 2026 traditional search engine query volume would drop by 25% in favor of AI assistants.

How ChatGPT chooses sources

ChatGPT draws on two layers of knowledge:

• Training knowledge. What the model learned up to its cutoff date. It has no citations and no fresh data. What matters is whether your brand and facts about it appeared on the web at all.
• Live search. When a question needs current information, ChatGPT searches the web, reads several pages and writes an answer with links to them. That's where the competition for citations happens.

Pages are added to ChatGPT's search by the OAI-SearchBot crawler. OpenAI keeps it separate from GPTBot, which collects content for training models, so you can allow the first and block the second. A third agent, ChatGPT-User, visits pages only at a user's request and doesn't decide whether you appear in results. ChatGPT has also used the Bing index from the start, so indexing in Bing still matters - details in the post on Bing, Copilot and IndexNow.

How Perplexity chooses sources

Perplexity searches the web live for every question. It works in four steps:

1. 1.It understands the question and often splits it into several sub-questions.
2. 2.It searches by keywords and meaning at the same time.
3. 3.It ranks the results for relevance and credibility.
4. 4.It writes an answer with numbered footnotes to the pages it actually used.

Pages are added to its index by the PerplexityBot crawler, while Perplexity-User fetches pages at a user's request. Perplexity clearly favors fresh, well-structured content from credible domains.

How ChatGPT and Perplexity differ

The data on most cited domains comes from a Profound analysis of 680 million citations from August 2024 to June 2025. It shows the difference well: ChatGPT leans toward encyclopedic sources, while Perplexity leans toward user conversations. I cover how a brand can show up in communities without getting banned in the post on Reddit, forums and UGC.

What increases your chances of being cited

The best data comes from the GEO-bench study (the paper "GEO: Generative Engine Optimization", KDD 2024). The most effective changes were adding citations to sources, expert quotes and statistics. They raised visibility by up to about 40%, while keyword stuffing didn't help. These are results from a controlled test, not from live ChatGPT, so treat them as a direction.

On top of that come signals confirmed in practice:

I show in detail how to write fragments engines like to cite in the post on how to write content AI cites.

Step by step - preparing your site

1. 1.Let AI search crawlers in. In robots.txt, allow OAI-SearchBot and PerplexityBot. You can block GPTBot if you don't want to share content for training - it doesn't affect ChatGPT search. Details in the guide on AI crawlers.
2. 2.Make sure you're indexed in Bing and Google. Submit your sitemap in Bing Webmaster Tools, not just Search Console.
3. 3.Remove technical barriers. Content should be visible without logging in and without waiting for JavaScript. More in the post on JavaScript SEO.
4. 4.Rewrite key pages for citation. An answer in the first two sentences, question-style headings, tables and an FAQ.
5. 5.Add data and sources. Every important figure with its origin.
6. 6.Build your brand entity. Consistent company details, an author with visible experience, and structured data linking to your profiles.
7. 7.Earn mentions where the engines cite from. Industry media, rankings, substantive forum threads.
8. 8.Update and date your content. Refresh the data in your key articles and show the update date.

A one-hour test kit

Before you change anything, check where you stand. Prepare 15 questions in three groups and ask them in ChatGPT with search turned on and in Perplexity:

For each answer, note three things: whether the brand appeared, whether your site is among the sources, and which domains the engine cites instead of you. That last list is your action map - it shows where it's worth earning a mention or what content to prepare. Repeat the test every month with the same questions. I describe an extended version that measures your share of answers in the post on the AI visibility audit.

The most common mistakes

• OAI-SearchBot or PerplexityBot blocked - often by accident, when blocking crawlers that train models.
• The site isn't in Bing's index - many companies only check Google.
• Text without specifics - there's nothing to cite.
• Content behind a login or loaded only through JavaScript.
• Inconsistent company details in different places on the web.
• Counting on llms.txt - neither engine has confirmed using it as a signal. More in the post on llms.txt.

---

I help brands show up in AI answers - from an LLM citation strategy and optimization for Perplexity and ChatGPT to full AI optimization (GEO). I also teach it in the SEO & GEO course. Get in touch - I'll check whether ChatGPT and Perplexity already cite you and what to change so they start.

Worth reading next:

• GEO (Generative Engine Optimization) - What It Is, How It Differs from SEO and Where to Start
• AI Visibility Audit - Does ChatGPT Recommend Your Brand?
• Bing, Microsoft Copilot and IndexNow - the Forgotten Ecosystem Feeding AI Answers
• Reddit, Forums and UGC - Why AI Models Cite Communities and How a Brand Can Show Up There