2026-08-03 · Jack Stovell
How to get cited by Perplexity
Perplexity is the most directly influenceable of the AI assistants, because it is retrieval-first: every answer is built from live search results, and it cites its sources by design. There is no waiting on training runs — the question is whether your page is in the results it retrieves, and whether it is the clearest answer among them. The work is making both true.
How does Perplexity choose its sources?
It searches first and writes second. When you ask Perplexity a question, it runs web searches, retrieves a ranked set of results, reads the pages it selects, and composes an answer with citations pointing back at them. The citation is not a courtesy feature — the cited answer is the product. That architecture is what makes it more influenceable than weights-first assistants: change what its search retrieves, and you change the answer.
Perplexity does not publish its ranking internals, so treat any confident recipe for them as guesswork. What is safe to say is what it documents — its crawler is PerplexityBot, and answers are composed from retrieved web sources — plus what is observable from using it: it favours pages that answer the question directly, and its citations change as the web does.
What gets a page into Perplexity's results?
The gate has four bars: crawlable, discoverable, direct, and current.
- Crawlable by PerplexityBot. Check that your robots.txt doesn't block it and that your CDN's bot defences aren't rejecting it at the edge — the AI crawler user agents list has the strings to look for. Like most AI crawlers it fetches raw HTML and does not execute JavaScript, so server-rendered content is a requirement, not a nicety.
- Discoverable. A sitemap, sane internal links, and inbound links from pages that already rank. Perplexity leans on search, so pages search can't find, it can't cite.
- Direct. Question-shaped headings whose first sentence answers the question, in self-contained passages an engine can lift whole. A page that makes the reader assemble the answer gives the engine nothing to quote.
- Current. Perplexity appears to lean towards recently updated pages for many queries, and visible dates on the page help it judge. We say "appears" deliberately — this is observation, not documentation.
None of this guarantees a citation. Being retrievable puts you in the pool; ranking decides who gets drawn from it, and how AI decides which sources to cite covers those deciding factors.
How do you check?
Ask Perplexity your target questions and read the citations — they are right there in the answer. Write down the five or ten queries you most want to be cited for, ask them, and record which domains it cites. Then study those pages: what do they do that yours doesn't? Usually the answer is mundane — they answer faster, carry a date, or sit on a site with more inbound links.
Re-test after you make changes, and re-test on a schedule regardless: answers shift as Perplexity's index updates, so a citation is a reading, not a rank you hold. The related and follow-up questions it offers are worth recording too — they are a free map of the query space as Perplexity sees it, and often surface adjacent questions your page should be answering. For the technical bar, the free AI visibility checker confirms whether your site is readable by AI crawlers at all.
The honest limit: a page nobody links to, on a topic where established sources already answer well, will struggle no matter how well-shaped it is. Perplexity rewards the same authority the rest of the web does — it just pays out in citations instead of clicks.
Check your AI visibility — free
Related guides
- How AI assistants decide which sources to cite
- Generative engine optimisation (GEO), explained
- AI crawler user agents: the full list (GPTBot, ClaudeBot, PerplexityBot and more)