How Perplexity Picks Its Sources and How to Get Your Brand Cited

Perplexity searches the internet in real time when someone asks it a question, according to its Help Center explainer dated August 2026. OnlyAEO is an AEO and GEO agency, and the steps below are the ones we run for Perplexity.

How does Perplexity decide which pages to cite?

Perplexity's choice of pages depends partly on the search mode the user picks, because the modes draw on different numbers of sources.

• Standard search. Perplexity's Pro Search help page describes standard search as surface-level, drawing on 1 to 2 sources.
• Pro Search. It runs multiple searches and draws on articles, academic papers, forums, videos and more, depending on the focus the user selects, such as Web, Academic, Finance or Files.

A standard search has few slots, so your page has to answer the exact question better than the alternatives. Pro Search can draw on forums and videos, which means what others say about you there counts too.

On ranking, Perplexity's Instant Buy help page says the answer engine ranks sources based on authority and relevance, and ranks product listings with similar criteria.

Can Perplexity's crawlers reach your site?

Only if your firewall lets both of Perplexity's agents through and your robots.txt allows PerplexityBot. The two agents follow different rules, so check access before you touch any content.

Perplexity's crawler documentation says PerplexityBot is designed to surface and link websites in Perplexity's search results. Perplexity-User is the fetcher that might visit a page when a user asks a question, and Perplexity may link that page in its answer.

Read the second row with the third: a firewall rule is the more likely reason a live fetch fails on your site.

How robots.txt groups match a user agent, and how to test a rule before you ship it, is covered in our robots.txt guide for AI crawlers. Perplexity's robots.txt help page adds that it disabled a feature that once let users ask it to summarize a specific URL blocked by robots.txt.

How we tested: CRM vendors' robots.txt files

To see how software companies treat Perplexity today, we read the public robots.txt files of well-known CRM vendors: HubSpot, Salesforce, Pipedrive, Zoho, Attio, Close and monday.com.

• Method. We fetched each file with curl in mid-October 2026 and read every group that applies to PerplexityBot and Perplexity-User.
• Headline result. None of the files we read blocks PerplexityBot from the whole site.
• Attio and monday.com name it. Attio gives PerplexityBot its own group that allows the whole site. monday.com lists PerplexityBot and Perplexity-User with the same disallow list it uses for its other AI bots.
• The rest use the wildcard. HubSpot, Salesforce, Pipedrive, Zoho and Close never mention Perplexity, so PerplexityBot follows their wildcard groups, which allow most or all of each site.
• Our own file. onlyaeo.com allows PerplexityBot and Perplexity-User by name.
• Limits. We read robots.txt only and sent no requests pretending to be Perplexity, because a firewall that verifies Perplexity's IP ranges would treat a spoofed request differently.
• No answer capture. A signed-out search on perplexity.ai hit a sign-in wall before any answer appeared, so we make no claim from it.

So a wildcard group is enough for PerplexityBot, as long as the pages you want cited are not on its disallow list. Name the agents when you want rules for them that differ from your default.

What makes a page easy for Perplexity to quote?

A page is easy to quote when the passage that answers the question can be lifted on its own and still make sense.

The GEO paper by Aggarwal and colleagues, accepted to KDD 2024, reports that its methods can boost visibility by up to 40% in generative engine responses. The same paper found that the effect of its strategies varies across domains.

The paper's page does not say the methods were tested on Perplexity, so read it as a result about generative engines in general.

These are the rules we write by, and this page follows them:

1. Answer under the heading. Put a short answer directly below each question heading, then the detail and the proof.
2. Make sections stand alone. Name the product and the subject instead of writing "it" or "this", so a quoted paragraph still reads correctly.
3. Keep each figure with its source. Write the number, who measured it and when in one sentence.
4. Use tables for real comparisons. Plans, specs and options in rows quote more cleanly than the same facts spread across paragraphs.
5. Phrase headings the way buyers ask. Use the wording from sales calls and search queries, then answer that exact question.

Heading order, passage length and tables are covered in more depth in how to structure content AI engines cite.

Does Perplexity prefer fresh content?

Less than most AI assistants in Ahrefs' study: Perplexity's in-text citations showed the next strongest preference for older content, after Google AI Overviews.

Ahrefs' freshness study, published in July 2025, analyzed 17 million citations across 7 AI search platforms, Perplexity among them. URLs cited by AI assistants averaged 1064 days since publication, against 1432 days for organic search results. Ahrefs puts that at 25.7% fresher on average.

Perplexity's in-text citations averaged 1166 days since publication. Ahrefs also saw Perplexity's in-text references seemingly ordered from newest to oldest.

One reading of those findings: recency may affect where you appear in an answer more than whether you appear. Ahrefs itself cautions that freshness is one of many factors, and that low-quality, irrelevant content updated every day will not have a magic positive effect on AI visibility.

Our advice: update a page when its facts change, such as a new price, feature or figure. We don't recommend changing the date on a page whose content is the same.

Which other websites does Perplexity cite, and how do you get on them?

Perplexity cites other websites as well as yours, and it marks some cited domains as Government, Academic or Trusted, according to its help page on source labels, dated October 2026.

The source labels page also explains how a site earns a label.

• Whole sites are rated. Perplexity rates the whole website a source comes from, not each individual page.
• The review asks plain questions. Does the site correct its mistakes, say who wrote each piece and keep news separate from advertising and opinion?
• Trusted reflects presence. The label covers sites that appear often enough in results and publish within their own area of expertise.
• Labels are not for sale. In Perplexity's words: "Labels are set by the source review process alone. Perplexity's partnerships, payments, and other business arrangements do not affect a site's label."
• No label is normal. Most domains have none, and Perplexity says that is not a negative judgment.

Check your own site against those questions first. Then get covered by the sites already cited for your buyers' questions.

Our earn-citation list does this for each client: it ranks the third-party sites AI cites in the brand's space where the brand is not mentioned yet, with a recommended play per site (guest post, partnership, listing or review). Our guide to earned media sources AI cites covers how to pitch them.

Some sources never touch the open web. Perplexity's premium data sources give users licensed content from publishers and data providers, searchable inside Perplexity with full citations. If the reference data in your category sits with such a provider, being in that data set may be another route into answers.

How do e-commerce brands get products into Perplexity answers?

Publish complete product data on pages Perplexity can reach. Product listings are ranked with criteria similar to the authority and relevance test described above, and the Instant Buy help page, dated August 2026, adds more detail.

• Detail helps. Perplexity is more likely to recommend merchants that provide deeper product details such as availability, reviews, pricing and specifications.
• No paid placement. Advertisers cannot pay to list their products in Perplexity's related products section, the Instant Buy help page says.
• Checkout has its own requirement. Instant Buy only works with merchants compatible with Perplexity's agentic checkout.

For a store, that puts product pages ahead of blog posts. Keep stock status, price, specs and reviews on the product page in plain text.

To track how products show up across assistants, see our guide to measuring AI visibility for e-commerce.

How do you measure whether Perplexity cites you?

Ask Perplexity your buyers' questions and read the sources it lists: the Help Center explainer says each answer includes numbered citations linking to the original sources. Then confirm from your server.

1. List the questions buyers ask. Use the same buyer wording as your headings (rule 5 above), grouped by product and by buyer role.
2. Run them and record the sources. For each answer, note the cited URLs, whether your brand is named and whether it is recommended. We count the answers that recommend a brand separately from those that only mention it in text or only link to it as a source.
3. Check your logs. Look for PerplexityBot and Perplexity-User requests on the pages you want cited. A Perplexity-User request suggests a user question led Perplexity to fetch that page.
4. Repeat on a schedule. Answers change, so rerun the same questions with the same settings and compare like with like.

Our crawler tracking recognises AI bots, logs which pages they read and captures AI-referred visits that GA4 misses. Our AI visibility platform can run ChatGPT, Claude, Gemini, DeepSeek and Grok. Perplexity is not among them, so for Perplexity we work from its visible citations and the crawler logs.

How long does it take to get cited by Perplexity?

Access problems show up fast, and citation gains take longer. Plan on at least 60 days before the first measurable results, which is what we usually see. No one can promise Perplexity citations: OnlyAEO does not offer a results guarantee and does not promise specific rankings or citations.

Each fix runs at its own speed. Perplexity says changes to its robots.txt settings may take up to 24 hours to show in its systems. A firewall fix can be tested as soon as it is live, because Perplexity-User may fetch pages when users ask.

Content and third-party coverage are the slow part. Our AEO timeline by engine compares Perplexity with the other assistants.

How to choose a partner for Perplexity work

Choose a partner that starts with access and evidence, and that tells you plainly what it cannot promise. These checks sort them quickly.

• Access first. Ask whether they check the items in the access table above before any writing.
• Sources on every claim. Every claim in the content they publish should trace to a named source.
• Measured engines. Ask which engines their software measures and how they check the rest, including whether they count recommendations apart from mentions.
• Engine coverage. Check that they work on the engines your buyers use. OnlyAEO works on ChatGPT, Claude, Gemini, Perplexity, Google AI Overviews, Microsoft Copilot, Grok, DeepSeek, Meta AI (Muse) and Instinct.

Most of the work carries across engines, and our guide to getting cited by AI engines covers what is shared.

More questions about getting cited by Perplexity

Does blocking PerplexityBot remove my pages from Perplexity?

Not entirely. Perplexity's robots.txt help page says PerplexityBot will not index the text of a site that disallows it, but Perplexity may still index the domain, headline and a brief factual summary of a blocked page.

Does Perplexity use my content to train AI models?

Perplexity says it does not build foundation models, so content it indexes is not used for AI model pre-training. Its crawler documentation adds that PerplexityBot is not used to crawl content for AI foundation models.

Do other crawlers feed Perplexity's index?

Yes. Perplexity partners with third-party crawlers to help build its search index and says it updated its agreements so they also respect robots.txt, particularly for news publisher sites.

How does Deep Research differ from a normal Perplexity answer?

Perplexity's Help Center says Deep Research performs dozens of searches, reads hundreds of sources and delivers a report in 2 to 4 minutes.