How does Perplexity choose sources?
Perplexity searches the web for each question, reads the top results and writes an answer with numbered citations, usually several per answer. The sources come from Perplexity's own index, built by PerplexityBot, plus pages fetched live when a question needs them.
The sources it picks tend to share four traits: they answer the question directly, they are current, they come from sites the index trusts for the topic, and they are easy to read in plain HTML.
What do PerplexityBot and Perplexity-User do?
| User agent | What it does | Respects robots.txt? |
|---|---|---|
| PerplexityBot | Surfaces and links websites in Perplexity's answers; not used for training foundation models | Yes |
| Perplexity-User | Fetches a page when a user's question needs it | Generally no, as the request comes from a person |
For AEO, allow PerplexityBot. Check your CDN and firewall too: bot protection on some hosts blocks it by default.
How do you get cited by Perplexity?
- Allow PerplexityBot in robots.txt and at the firewall.
- Answer first. Open each section with a 40 to 60 word answer to its heading.
- Add numbers with sources. The GEO research tested Perplexity directly and found adding statistics, quotations and citations raised a source's visibility by up to 37% there.
- Keep pages current. Show a visible update date and refresh facts on a schedule.
- Keep the HTML clean. Content should be readable without running scripts.
- Cover the follow-ups. Perplexity suggests related questions after every answer; those are fan-out queries you can answer on the same page.
How do you track Perplexity citations?
Ask Perplexity your key questions every month and log the cited URLs. Perplexity also sends referral traffic you can see in analytics under its domain. Track which landing pages get those visits; they show which content Perplexity trusts. How to measure AEO covers a monthly tracking routine for all AI tools at once.