Perplexity cites more openly than any major assistant, which makes absence easy to diagnose. Usually it is one of three things: PerplexityBot cannot reach you, your content is older than the sources it prefers, or the answer is being taken from forums and community threads rather than company websites.

The short version

Why this happens

Perplexity is the most transparent of the major assistants, and that is a gift for diagnosis. Every answer lists its sources. You are not guessing at why you were omitted, you can read what was used instead.

Recency weighting. Perplexity favours fresh material more visibly than the others. A comprehensive guide from 2023 will lose to a thinner but current page on the same subject far more often here than in a Google AI Overview. If your best content is old and undated, that is frequently the whole explanation.

Community sources. This is the one that surprises people. Perplexity draws heavily on Reddit, specialist forums and discussion threads, because they contain first-hand experience in the language people actually use. For a great many practical questions, a forum thread genuinely is a better source than any company’s marketing page, and Perplexity treats it that way. If your category is discussed in communities where you have no presence, you are absent from the conversation that shapes the answer.

Crawler access. Two agents again, governed separately. PerplexityBot builds the index. Perplexity-User fetches a page because a user’s question required it. Blocking either has different consequences, and firms sometimes block one while assuming they have allowed everything.

Answer shape. Perplexity composes short, direct answers with citations attached. Content that answers in one clean paragraph is structurally advantaged over content that builds an argument across a page.

How to check it yourself

  1. Ask the buying question your best customer would ask, and read the source list. Those are the pages you are competing with, named explicitly.
  2. Note the dates on the cited sources. If everything cited is from the last few months and your equivalent page is two years old and undated, you have your answer.
  3. Count how many cited sources are forums rather than company sites. In many categories it is a large share.
  4. Check robots.txt for both PerplexityBot and Perplexity-User.
  5. Ask a follow-up question. Perplexity will re-retrieve, and watching which sources persist across follow-ups tells you which are considered strongest.

What to do about it

  1. Allow both agents and confirm nothing upstream is blocking them.
  2. Date your content visibly and keep the important pages genuinely current. Not a changed timestamp, a changed page.
  3. Answer in the first paragraph, cleanly, with the qualification after rather than before.
  4. Show up genuinely in the communities where your category is discussed, under a real name, over time. Not astroturfing, which communities detect faster than models do.
  5. Use Perplexity as your early indicator. Because it moves fastest, it usually shows improvement before the others do, which is useful for keeping a programme funded.

The full method is in how to appear in Perplexity and optimising for Perplexity.

Platform behaviour changes. This page reflects how Perplexity worked in August 2026 and is reviewed quarterly.

Back to all GEO questions →