How to rank in Perplexity

Perplexity is built as an answer engine: it rewrites your question into searches, retrieves pages from its own index and other sources, reads them, and writes a short answer with numbered citations. Unlike ChatGPT, it searches by default, so almost every answer has sources you can inspect.

Perplexity documents two agents. PerplexityBot "is designed to surface and link websites in search results on Perplexity" and, per Perplexity, is not used to crawl content for AI foundation models. Perplexity-User fetches a page when a user's question needs it; because the user asked, Perplexity says it generally ignores robots.txt.

If robots.txt or your firewall blocks PerplexityBot, your pages will not be in Perplexity's index. Perplexity recommends allowing it:

User-agent: PerplexityBot
Allow: /

Check CDN-level bot blocking too; see AI crawlers and robots.txt.

Perplexity's answers are short and dense with facts, and it cites the passages it uses. Pages that do well:

• answer one question per section, directly, in the first sentences;
• carry specific, checkable facts: numbers, dates, prices, names;
• are recently updated, with the date visible, since many Perplexity questions are time-sensitive;
• use descriptive headings and tables that can be lifted out cleanly.

See how to write content AI assistants cite.

Run your key questions in Perplexity and look at the sources panel. You will usually see a mix of news sites, Reddit, YouTube, review sites, Wikipedia and a few niche blogs. Those are your targets: get mentioned in them, or publish something better than them.

News pages are especially common in Perplexity answers about companies and launches, which is one reason a press release on news sites often shows up as a Perplexity source shortly after it runs.

Perplexity sends referral traffic from perplexity.ai, visible in analytics. Combine that with a monthly check of your question set; see how to track AI traffic in Google Analytics.