How Does Perplexity Choose Its Sources?

最終更新 2026-07-03

Perplexity searches the live web for every question, ranks the results, and writes an answer from the top pages, citing them inline. It favors pages that directly and clearly answer the exact query, come from credible domains, and are fresh. The citations are the sources it actually read.

Retrieval first, then writing

Unlike a model answering from memory, Perplexity is retrieval-augmented: it fetches pages in real time, then grounds its answer in them. That makes citations central. If your page is not retrieved, you cannot be cited, and if you are not cited, you are effectively invisible in that answer.

What makes a page citable

  • It answers the specific question directly, high on the page, in plain language
  • The domain has topical authority and independent trust signals
  • The content is current and dated, not stale
  • The claim a reader wants is a clean, quotable sentence, not buried in fluff
  • The page is crawlable, fast, and not blocked to bots

How to earn Perplexity citations

Write the liftable sentence. For every question your buyers ask, publish a page that answers it in the first paragraph, backs it with specifics, and earns links from sources Perplexity already trusts. CitationWorks shows which of your pages get cited, which competitors get cited instead, and where your trust gap is widest.

アクセスを申請

AIが今日あなたのブランドをどう語っているか、何を最初に直すべきかを確認できます。

  • 01現在のAI上での露出
  • 02正確性とリスクのギャップ
  • 03専任エキスパートのガイド

通常24時間以内に返信。自動フォローアップはありません。

送信することで、プライバシーポリシーに同意したことになります