Perplexity’s help centre says “Every answer includes citations linking to the original sources”. That one sentence makes Perplexity the assistant where a B2B page can watch itself being cited or not. There is no answer without a source list. If you are in it, you can see it. If not, you can see who is.
That’s why Perplexity’s source list is the first one we read in the free check. The 3 conditions below are what we look at on a page before we do anything else.
Every answer starts with a live search
The same help article, last modified 3 September 2026, says Perplexity “searches the internet, gathering information from authoritative sources like articles, websites, and journals”. Then it “compiles the most relevant insights into a coherent, easy-to-understand answer”. The models are named on the page. GPT-5 and Claude 4.6 Sonnet are among them.
For your page, the order of events is fixed. A search runs and pages come back. A model writes from them and numbers the sources. A page the search doesn’t return is not in the answer, whatever the model remembers about you. So the first question is plain. Can the search find the page at all?
Two agents, and both have to get through
Perplexity’s crawler documentation names them. The first: “PerplexityBot is designed to surface and link websites in search results on Perplexity”. The page adds that it “is not used to crawl content for AI foundation models”. The second is Perplexity-User. “When users ask Perplexity a question, it might visit a web page to help provide an accurate answer and include a link to the page in its response”.
Allow both by name in your robots.txt, the text file that tells crawlers what they may read. Ours does. They sit in the same group as the OpenAI and Anthropic agents. The documentation says changes can take “up to 24 hours” to take effect, so a fix made on Monday is checked on Tuesday.
Then the firewall. Perplexity’s page has a section on it. “If you’re using a Web Application Firewall (WAF) to protect your site, you may need to explicitly whitelist Perplexity’s bots to ensure they can access your content”.
It gives the rule for 2 common firewalls: a rule that allows requests matching both the agent’s name and Perplexity’s published addresses, placed above your blocking rules. A CDN is the service in front of your server that also filters bots. If your site sits behind one, this is where a citation gets lost. Robots.txt shows nothing.
The site is rated as a whole, and 3 questions decide it
Perplexity publishes what it looks at when it rates a source, in an article last modified 9 September 2026. The rating is per site: it “rates the whole website a source comes from, not each individual page”. The tests are whether the site corrects its mistakes, whether it says who wrote each piece, and whether it keeps “news separate from advertising and opinion”.
Two things on that page are worth reading before anyone sells you a label. “Most domains on the web do not have a label”, and no label “is not a negative judgment”. And: “Perplexity’s partnerships, payments, and other business arrangements do not affect a site’s label”.
So there is nothing to buy. There are 3 habits.
Here is where we stand on them. Our own site is the specimen. Corrections: a factual error on our pages is fixed when it is found, and the terms say factual errors in client work “are fixed at no cost and without limit”. Authorship: our posts carry the studio’s name rather than a person’s. That is a choice we made, and a test we only half pass. Separation: there is no advertising on the site to separate.
One and a half out of 3, stated plainly. The page says it rates the whole site, and a claim that hides its gaps is the thing being rated.
The passage has to survive being numbered
A Perplexity citation is a number in the text pointing at a page. Your reader clicks the number, lands on the page, and looks for the sentence the answer was built on. Our anatomy of a quotable page has the 5 tests. The 2 that matter most here are the opening paragraph and the date.
The opening paragraph, because that’s the passage most likely to be lifted, and it has to answer the buyer’s question with the number in it. The date, because Perplexity’s help centre tells users to “double-check sources for added confidence”, and a page with no date gives a checking reader nothing to check.
What we’d do on Monday morning
Open robots.txt and confirm PerplexityBot and Perplexity-User are allowed by name. Ask whoever runs the CDN whether either is blocked there, and get the answer as a log line rather than an assurance.
Then take the page that should carry the citation and read its opening paragraph as a stranger would. Add a visible date if it hasn’t got one. Then ask Perplexity your 3 buyer questions. Read the source list.
That last step is the free check, and we’ll do it with ChatGPT and Gemini alongside. If the page that should carry the citation doesn’t exist yet, the AI citations article comes with the Perplexity and ChatGPT source lists read again 30 days after delivery.