How to get cited by Perplexity: a guide for B2B SaaS teams

Perplexity runs a live web search for each question and numbers the sources behind its answer, so getting cited depends on three things you can influence. Its crawler must be able to reach your pages. One of those pages must answer the buyer's question in plain text. And the third-party pages it already numbers for your category, such as reviews, roundups and forum threads, should mention you. For a B2B SaaS team that becomes four jobs: let PerplexityBot through robots.txt and the firewall, publish pricing, comparison, alternatives and integration pages with their facts in the HTML, keep the dates on those pages honest, and earn a place on the sources Perplexity keeps returning to.

Book a 30-minute call if you would rather see your starting point first: we will look at your category live and show you what the engines say today when a buyer asks for a recommendation.

Every fact below about how Perplexity behaves comes from its own documentation and announcements, read on 16 October 2026. None of those pages explains how Perplexity chooses between candidate pages, so treat any published list of "Perplexity ranking factors" as somebody's inference, and prefer advice that rests on something you can check.

How Perplexity turns a question into numbered sources

Perplexity's help centre lays out the sequence: it works out what the question is asking, searches the web at that moment, and writes a short answer from what it found. The citations are part of the format, not an extra. In Perplexity's own words, "each answer includes numbered citations linking to the original sources" (How does Perplexity work?). Its Pro Search mode goes wider: it runs several searches across articles, academic papers, forums and videos, and can draw on dozens of sources for a single answer (What is Pro Search?).

Two consequences matter for a software company.

You compete question by question. Each answer numbers a limited set of sources, and the pages best placed to earn a number are the ones that answer that specific question outright. A homepage that gestures at five use cases gives an answer nothing to number.

The search runs when the question is asked. A correction can reach answers once Perplexity has fetched the new version of the page, without waiting for any model to be retrained. That cuts both ways: a stale price on your pricing page is repeated just as readily.

If your team is still weighing how far this departs from ordinary SEO, AEO vs SEO: what actually changes covers the move from ranking pages to being chosen as a source.

Perplexity's two agents, and what robots.txt does to each

Perplexity's crawler documentation lists two user agents. They do different jobs, and they treat robots.txt differently.

	PerplexityBot	Perplexity-User
Its job	Crawls the web so pages can appear, with links, in Perplexity's search results	Fetches a page when a user's question calls for it, and may link that page in the answer
robots.txt	Perplexity recommends allowing it, along with its published IP ranges, if you want to appear in results	Generally ignores robots.txt, since a person asked for the fetch
Training	Perplexity says this crawler does not feed AI foundation models	Perplexity says it neither crawls the web nor gathers training data
IP list	perplexity.com/perplexitybot.json	perplexity.com/perplexity-user.json

Source: Perplexity Crawlers, Perplexity's documentation, read 16 October 2026.

On the second agent, the documentation is blunt: "Since a user requested the fetch, this fetcher generally ignores robots.txt rules." What that means in practice:

Allow PerplexityBot. It is the crawler that brings your pages into Perplexity's search results. A rule that blocks it, whether a group that names it or a catch-all User-agent: * group that disallows everything, withholds the access Perplexity recommends for any site that wants to appear. Perplexity-User may still fetch a blocked page when a question calls for it, but a page kept out of the crawl is not a strategy.

Know which group it reads. Under the robots.txt standard, RFC 9309, a crawler follows the group that names it and falls back to the User-agent: * group only when no group names it. So the moment you add a PerplexityBot group, the disallows in your catch-all group stop applying to it. Repeat any paths you still want left alone:

User-agent: *
Disallow: /app/
Disallow: /admin/

User-agent: PerplexityBot
Disallow: /app/
Disallow: /admin/

Give a change a day. Perplexity says an edit to robots.txt can take up to 24 hours to take effect.

Check every host, not only www. The same standard puts the file at the root of each host, so a SaaS company's www, docs, help and app subdomains each answer for themselves. Plan limits, integration lists and security details often sit on the docs or help host, and a file left over from a migration can hide them.

Look past robots.txt to the firewall. Permission in robots.txt means nothing if a firewall or bot-protection rule turns the request away. Perplexity's documentation gives allow-rules for Cloudflare and AWS WAF that match the user agent and the published IP ranges together. Ask whoever runs your CDN to check its bot settings.

Confirm the visits in your logs. Search your server or CDN logs for both user agents and check the source addresses against the two IP lists, because anyone can send a request that calls itself PerplexityBot. If PerplexityBot never shows up, fix that before you write another page.

Assume no JavaScript. Perplexity's crawler documentation does not say whether either agent runs JavaScript. Until it does, put every fact you want quoted in the HTML your server sends: prices, plan limits, integrations, security reports, comparison tables. A pricing calculator that only renders in the browser may be invisible to an agent that does not run scripts.

Pages that answer a buyer's question

The questions worth winning are the ones asked close to a decision, and each kind has a page type that answers it best.

The buyer asks	The page that answers it	What it has to state as text
Best [category] tool for a [segment] team	Use-case or category page	Who it is for, the job it does, where it stops
Alternatives to [rival]	Alternatives page	Why teams switch, what you do differently, what moving involves
[Your product] vs [rival]	Comparison page	A side-by-side table, with dated facts for both products
How much does [your product] cost?	Pricing page	Each plan's price, what it includes, the billing terms
Does [your product] work with [tool]?	Integration page	What syncs, in which direction, on which plans
Is [your product] SOC 2 compliant?	Security or trust page	Which reports exist, how recent they are, how to request them

Writing these so an answer can number them is mostly discipline:

• Answer under the heading. Put the direct answer in the first sentence or two and the detail after it. An answer engine needs a sentence it can lift without interpreting three paragraphs first.
• One page, one question. A comparison page that is also trying to be a pricing page answers neither cleanly.
• Specifics over adjectives. "SSO is included from the second plan up" can be repeated; "enterprise-grade security" carries nothing worth repeating.
• Date and source your rival facts. Prices and features change. Link each rival fact to where you read it and say when, so a reader, or an engine, can check it.

How to structure the two page types that carry most of this, alternatives and head-to-head comparisons, is covered in SaaS alternatives and comparison pages.

Dates Perplexity can read

Perplexity also sells a Search API to developers, and its documentation shows what that search returns for each page. Each result comes back with fields for its publication date and its last-updated date, and developers can filter results by either one, or by recency windows that run from the past hour to the past year (Search API date filters).

That is the developer product, not the answer engine, and Perplexity does not say how much recency counts in a consumer answer. What it does show is that Perplexity's search keeps track of when a page was published and when it last changed. Give it accurate values:

• Show a visible "last updated" date on any page that carries facts, and keep dateModified in the page's structured data and lastmod in the sitemap in step with it.
• Move the date only when the content changes. A new date on an unchanged page is noise.
• Put pricing, integration and comparison pages on a review schedule. They hold the facts buyers ask about, and they go out of date first.

Earn a place on the pages Perplexity already numbers

Your own site is only part of the evidence. Run your ten most important buying questions in Perplexity and write down every numbered source. Sort them into three piles: your pages, rivals' pages, and third-party pages such as review sites, roundups, trade publications and community threads. The third-party pages that come up again and again are your citation targets, because they are where Perplexity goes for your category today.

Getting onto them is slow, ordinary work. Give roundup editors the facts and a reason to include you. Keep review-site profiles complete and current. Take part in community threads openly as the vendor, where the community's rules allow it, with answers that help people who will never buy. How to get a SaaS product into best-software lists goes through the outreach step by step.

What the publishers' programme is, and what it is not

Perplexity has announced a programme for publishers, and it is easy to misread. The Perplexity Publishers' Program was announced on 30 July 2024 with six launch partners, among them TIME, Der Spiegel and Fortune. As announced, publishers would get a cut of whatever Perplexity earned from interactions that referenced their content. In August 2025 Perplexity announced Comet Plus, a $5 standalone subscription whose revenue would go to participating publishers after Perplexity kept a small share for compute, divided according to human visits, search citations and agent actions.

Both were announced as ways of paying publishers. Neither announcement says that joining changes which sources an answer cites. For a SaaS team, citations still come through the pages Perplexity already numbers.

Measuring your Perplexity citations

Perplexity is an easy engine to audit by hand, because every answer shows its numbered sources. Keep a fixed list of buying questions and run each one a few times on the same day each month. For every answer, record four things: whether you are named, where you appear, whether a page of yours is among the numbered sources, and which third-party pages are numbered.

Two first-party signals fill in the rest. Perplexity-User visits in your logs show live questions reaching your pages, since that agent acts when someone asks Perplexity something. Visits referred by perplexity.ai in your analytics show readers following a citation through to you. Neither is complete, but both are yours and both are free.

The full method, with scoring and share of answer against named rivals, is in how to measure brand visibility in ChatGPT and AI answers.

A first month, in order

1. Week one, access. Read robots.txt on every host, check the firewall's bot settings, and confirm PerplexityBot in the logs. This comes first because every later step assumes the crawler can reach you.
2. Week two, baseline. Run your buying questions, record who is named and which sources are numbered, and list every wrong fact about you.
3. Weeks three and four, pages. Fix the pages that answer your most valuable questions, starting with pricing and the comparison against the rival you meet most, with the facts in the HTML and the dates accurate.
4. From then on, sources. Work through the recurring third-party pages one at a time, and rerun the questions every month to see which changes are landing.

How this guide was researched

Every fact here about Perplexity's behaviour comes from Perplexity's own pages, read on 16 October 2026: the crawler documentation, the help-centre articles How does Perplexity work? and What is Pro Search?, the Search API documentation on date filters and result fields, and the announcements of the Publishers' Program (30 July 2024) and Comet Plus (25 August 2025). The robots.txt rules on hosts and groups come from RFC 9309, sections 2.2.1 and 2.3. Where Perplexity's documentation is silent, on how it ranks candidate pages and on JavaScript, this guide says so rather than guessing.

Getting help with this

The Rank Master gets B2B SaaS companies found in Google and recommended by the seven AI engines buyers ask first: ChatGPT, Perplexity, Gemini, Claude, Grok, Google AI Overviews and Google AI Mode. The work runs as three workstreams at once: revenue pages built for the questions asked at the end of an evaluation; earned citations on the listicles, review sites and publications those engines quote, never bought from a network; and open measurement in a live dashboard and monthly report. You approve every page before it publishes, and your pages, briefs, data and contacts stay yours.

Plans come with published pricing, a 90-day initial term, and then month-to-month on 30 days' notice. Book a 30-minute call: we will look at your category live, show you what the engines say when a buyer asks for a recommendation, and give you a straight answer on whether we can help.