To get your business mentioned by Perplexity, make sure both of its crawlers, PerplexityBot and Perplexity-User, can reach your site. Then publish specific, current pages on your own domain that answer the exact questions your buyers type into it, knowing that nobody can guarantee a mention.
Perplexity cites pages it can reach that answer a question more directly than the alternatives. The rest of this article covers how it finds pages, what it tends to cite, what we have seen in our own measurements, and how to tell whether any of it is working.
How does Perplexity find the pages it cites?
Perplexity runs its own crawlers. Its crawler documentation describes two user agents with different jobs:
- PerplexityBot is designed to surface and link websites in Perplexity's search results. Perplexity states it is not used to crawl content for AI foundation models.
- Perplexity-User visits pages on behalf of a user's specific query, to help generate an answer at that moment.
That gives you two routes in. PerplexityBot builds the pool of pages Perplexity can surface, its own index, separate from Google's. Perplexity-User reads a page live while a person is waiting for an answer.
| PerplexityBot | Perplexity-User | |
|---|---|---|
| Purpose (per Perplexity's docs) | Surface and link websites in search results | Fetch pages for a specific user query to help generate an answer |
| When it visits | Ahead of time, as it builds what Perplexity can surface | During a live answer, triggered by a real question |
| Used for training foundation models | No, according to Perplexity | Not its purpose; it fetches on behalf of a user |
| What you lose if you block it | Your pages are less likely to be among those Perplexity can surface | Perplexity cannot read your page at the moment it answers |
The live route shows up in server logs. In about two weeks of logs, an edtech platform we work with received 1,918 fetches on content pages from live retrieval agents (ChatGPT-User, OAI-SearchBot and Perplexity-User combined). A client blog of a few dozen articles received 56 such fetches across 24 different articles. Each fetch means an assistant read a page because a real person asked something.
Can Perplexity actually reach your site?
Check this before you write a single new page. If robots.txt, a firewall or a CDN bot rule turns these crawlers away, nothing else you do matters.
We got this wrong ourselves. On itgrows.ai, a robots.txt rule meant to hide one redirect page, /b, was written as a prefix. Robots.txt rules match prefixes, so from 11 August to 21 October 2026 that one line also closed /blog, /business and every article in our business blog to all crawlers. Bing Webmaster Tools had listed our business page as 'discovered but not crawled' since 11 August 2026. Google had none of the business pages indexed. We changed the rule to /b$ on 21 October 2026.
A practical check:
- Open yourdomain.com/robots.txt and read every Disallow line. Short rules are the dangerous ones, because they match every path that starts with them.
- Look for groups addressed to PerplexityBot, Perplexity-User or a blanket rule for all user agents.
- Ask whoever runs your CDN or firewall whether bot protection blocks or challenges AI crawlers.
- Search your server logs for both user agent names to confirm they actually arrive.
We go through this in more detail in Is My Website Blocking GPTBot, ClaudeBot and PerplexityBot?. Submitting a sitemap and using Bing Webmaster Tools and IndexNow helps new pages get discovered quickly. The Bing side is explained in Does ChatGPT Use Bing?.
One more warning: being indexed by Google does not mean assistants can find you. On 19 October 2026 we asked assistants for the exact titles of 18 articles, each older than three weeks, on three client sites. ChatGPT found 1 of 18 and Claude 7 of 18, while Google had indexed 29 of 30 articles of one of those sites. Perplexity has its own index, so treat it as its own channel.
What kind of page does Perplexity cite?
The short answer is pages that are specific, current and close to being a primary source. Three things we measured point the same way.
Primary sources win general questions
In a test on 23 September 2026 we put 20 German business questions to an assistant. It answered 10 without searching at all and cited nobody. Where it did search, it cited primary sources on 9 of 10: laws, ministries, the OECD, large accounting firms. Firms were named only on 'who should I hire' and 'how do I choose' questions.
The implication is uncomfortable. A generic blog post that rewrites what a ministry already says will lose to the ministry. Your page has a chance when it is the primary source for something only you can state: your prices, your process, your service area, your turnaround times, your own data, the specific situations you handle.
Your own website and directories both count
On 20 October 2026 we measured answers for an orthopaedic practice in Bonn: 20 patient questions, four assistants. The practice was named in 12 of 80 answers, and Perplexity named it in just 1 of 20. The directory jameda.de was cited in 12 answers. Competing practices' own websites were the most cited sources. 'Who should I choose' questions named practices; questions about costs, risks and treatments named none.
So two surfaces matter: a website that answers patients' and buyers' questions in detail, and accurate profiles on the directories assistants already cite in your field.
Freshness and specificity are how a page gets picked
When Perplexity searches, it chooses among the pages available for that query at that moment. A page with a clear date, current prices and a direct answer in its first lines is easier to quote than a broad overview from three years ago. Update pages when something actually changes, rather than just bumping the date.
Markup is not the lever
Google's documentation on AI features says you don't need new machine-readable files or special schema.org markup to appear, and its guide to generative AI search lists overfocusing on structured data as unnecessary. We saw the same pattern: a site with server rendering, Article, FAQPage and Organization JSON-LD and a sitemap, but only 75 URLs, was absent from answers on its own article's exact topic. A site with about 190,000 crawlable URLs was crawled constantly. More pages that genuinely answer real questions give crawlers more to find. Better tags on a handful of pages do not.
Our first client citation came from Perplexity
We measure assistants' answers every week, and we started with ourselves. On 21 October 2026 we put 20 English and 20 German questions about getting a business into AI answers to ChatGPT, Claude, Gemini and Perplexity, three times each. ItGrows was named in 0 of 160 answers, and none of our pages was cited. We took that measurement on the same day we corrected the robots.txt rule described above, so we treat it as our starting point. The full write-up is in We Checked Our Own AI Visibility and Were Named in 0 of 160 Answers.
The first citation we saw for a client came from Perplexity. In our weekly measurement of 20 October 2026, Perplexity cited an article we had published on a client's site, a German company-formation advisory.
One citation is not a trend, and we would not promise it repeats. But it fits the pattern: a specific article, on the client's own domain, answering a narrow question a founder would actually ask, on a site the crawlers could reach.
How do you know whether Perplexity mentions you?
Ask the same question three times and you will often get different businesses named. The 2023 to 2026 survey of GEO research reaches the same conclusion: engines cite different sources that vary across runs, so visibility has to be measured across several engines and over time. The paper that introduced GEO proposed visibility metrics for the same reason.
A workable method:
- Write down 20 questions your buyers really ask, including 'who should I choose' questions.
- Ask each one three times and count your business as named if any of the three asks names it.
- Record which sites are cited, because that tells you which pages and directories you are competing with.
- Compare weeks only on the same questions with the same number of asks.
Perplexity is cheap to measure. At API prices of October 2026, one question asked once with web search costs us about $0.0013 with Perplexity.
Do not use referral traffic as your metric. In about two weeks of logs from three of our course websites, Perplexity sent zero human visitors. Assistants mostly answer in place, so a mention can influence a buyer without producing a click.
To see where you stand today, the free check at itgrows.ai generates five buyer questions from your homepage and gets answers from ChatGPT and Perplexity in 25 to 40 seconds. If you want this done continuously, ItGrows publishes answer articles on your domain, fixes crawler access and indexing, and measures the four assistants weekly for $499 a month (plus a one-time $499 setup, free with 3 months or a year). For a manual method, see How to Check Whether AI Assistants Mention Your Business.
Which mistakes keep businesses out of Perplexity's answers?
- Prefix rules in robots.txt. A short Disallow line can close far more than you meant, as it did for us.
- Allowing one Perplexity agent but not the other. Allowing PerplexityBot while blocking Perplexity-User, or the reverse, closes one of the two routes.
- Generic content. Pages that restate what authorities already publish lose to the authorities.
- Only targeting informational questions. In our measurements, businesses were named mainly on 'who to choose' questions.
- Ignoring directories. If a directory dominates citations in your field, your profile there is part of your visibility.
- Trusting a single ask. One answer tells you almost nothing; repeated asks over weeks do.
Frequently Asked Questions
Does blocking PerplexityBot keep my content out of AI training?
According to Perplexity's documentation, PerplexityBot is not used to crawl content for AI foundation models. Its job is to surface and link websites in search results, so blocking it mainly reduces your chance of appearing in Perplexity's answers.
How long does it take before Perplexity mentions my business?
There is no fixed timeline, and anyone who promises one is guessing. Fix crawler access first, publish specific pages for your buyers' questions, then measure weekly on the same questions.
Do I need schema markup or special AI files for Perplexity?
Google states that no special markup or AI text files are needed for its AI features, and in our measurements a fully marked-up site with few pages still did not appear. Keep markup for rich results, but put your effort into crawlable, specific content.
Will being mentioned by Perplexity bring visitors to my website?
Possibly, but do not count on it. In about two weeks of logs from three of our course websites, Perplexity sent zero human visitors, because assistants usually answer in place. Track mentions and citations instead of referral traffic.