Perplexity SEO: how to get cited in Perplexity answers
Short answer Perplexity cites pages from its own web index, not Google's. First make sure its crawler, PerplexityBot, gets past your robots.txt and firewall. Then give it pages worth quoting: one page per service, plain facts in the first lines, the words customers use, and a complete Yelp page. Nothing guarantees a citation, so recheck monthly.
Want help applying this to your business? Talk it through on a free 30-minute call.
Book a callHow does Perplexity choose what to cite?
It searches its own index, ranks passages from the pages it finds, and writes an answer with numbered links to them. Ranking on Google doesn't carry over: Perplexity runs its own crawler and keeps its own index.
Perplexity's research team described that system in September 2025 (Perplexity Research: Architecting and evaluating an AI-first Search API). The index tracks over 200 billion URLs. Four of its steps tell you what to do with your pages:
| What Perplexity says it does | What it means for your page |
|---|---|
| A model decides when to re-read each page, by its importance and how often it changes | Update a page whenever its facts change, so the copy Perplexity holds isn't old |
| Pages are cut into passages, and each passage is ranked on its own | Every section must make sense read alone: name the service and the town in it |
| It searches by exact words and by meaning at the same time | Use the words customers type, not only your own terms |
| Early filters remove "clearly non-responsive or stale content" | Delete expired offers, old prices and last year's dates |
The passage step matters most:
"Having well-structured, comprehensive parsing allows us to decompose documents into self-contained spans, each of which can be individually retrieved and ranked at query time." (Perplexity Research)
Then each answer "includes numbered citations linking to the original sources" (Perplexity Help Center: How does Perplexity work?). Perplexity hasn't published how it weighs these steps, and nothing guarantees a citation.
Can Perplexity read your website?
Usually yes, unless a robots.txt rule, a firewall or a site-builder setting turns its crawler away. Perplexity's own advice:
"To ensure your site appears in search results, we recommend allowing PerplexityBot in your site’s robots.txt file and permitting requests from our published IP ranges listed below." (Perplexity Docs: Perplexity crawlers)
It sends two visitors (Perplexity Docs: Perplexity crawlers; Perplexity Help Center: How does Perplexity follow robots.txt?):
| Visitor | When it comes | Obeys robots.txt? |
|---|---|---|
| PerplexityBot | Ahead of time, to build the search index | Yes |
| Perplexity-User | When someone's question needs your page | It "generally ignores robots.txt rules" |
Check these four things, in order:
| Check | How | Fix |
|---|---|---|
| robots.txt | Open yoursite.com/robots.txt. Look for a block on PerplexityBot, or "User-agent: *" followed by "Disallow: /" | Remove the block. Perplexity says changes can take up to 24 hours |
| Firewall or CDN | Ask your host whether a bot filter is on. Perplexity gives setup steps for Cloudflare and AWS | Allow Perplexity's two published IP lists. On Cloudflare, Bot Fight Mode can't be skipped by a custom rule and can still challenge Perplexity's bots (Perplexity Docs) |
| Site-builder AI setting | Read the list of bots it blocks. Squarespace's "Block known artificial intelligence crawlers" box doesn't include PerplexityBot (Squarespace Help) | Keep search bots allowed |
| Server or host logs | Search for "PerplexityBot" | If it only gets errors or challenges, something above still blocks it. No visits at all can also mean it hasn't come yet |
A block also leaks. On a page that blocks PerplexityBot, Perplexity "may still index the domain, headline, and a brief factual summary" (Perplexity Help Center: How does Perplexity follow robots.txt?). So an accidental block can leave Perplexity with your name but not your details.
The full robots.txt and Cloudflare walkthrough, for every AI bot, is in should you block AI crawlers?
What makes a page worth citing?
A page that answers one question plainly, near the top, on a site that shows who stands behind it.
Perplexity marks some cited sites with a Government, Academic or Trusted label. Its review asks objective questions, such as whether a site corrects its mistakes, says who wrote each piece, and keeps news apart from advertising and opinion (Perplexity Help Center: Understanding source labels). A local business site is unlikely to carry a label, and Perplexity says that costs nothing:
"No label does not mean a site is low quality, and it is not a negative judgment." (Perplexity Help Center: Understanding source labels)
Perplexity doesn't say labels change which pages it cites. Its questions are still a fair bar for any page you want quoted.
| On the page | Why |
|---|---|
| First two sentences: the service, the town, who it's for | The top passage has to stand alone |
| One page per service, one question per section | Each passage is matched to one question |
| Headings in the customer's words: "water heater replacement", not "hot water solutions" | Perplexity matches exact words as well as meaning |
| The owner's or author's name, and the date the facts were last checked | Perplexity's source review asks who wrote each piece; a date shows the facts are current |
| Prices, service areas and hours in lists or tables | Perplexity says pages with well-structured lists and tables "may benefit from more formulaic parsing" (Perplexity Research) |
Your listings count too
Listing sites get cited as well, mostly while people are still choosing. Yelp reports that on ChatGPT, Gemini, Google AI Mode and Perplexity, 91% to 97% of its citations come on prompts that name no business, nearly 60% on prompts with "best" and 34% on prompts with "near me" (Yelp for Business, September 2026).
Perplexity's help center has no page for claiming a business listing, so your Yelp page and directory listings carry your hours, services and reviews. Keep your name, address and phone identical everywhere: local citations and directories.
A 30-day plan for Perplexity
| Days | Do this | Done when |
|---|---|---|
| 1–3 | Write 10–15 questions customers ask, like "best emergency plumber in Tacoma". Ask each in Perplexity. Log whether you're cited, which page, what it says, and who else is cited | You have a baseline sheet |
| 4–7 | Run the four access checks above | PerplexityBot shows in your logs, or your host confirms nothing blocks it |
| 8–14 | Rewrite the three service pages customers ask about most: plain first lines, customer words, one question per section, a date | Each page answers its question in two sentences |
| 15–21 | Claim and complete your Yelp page. Fix every wrong fact Perplexity repeated, where it lives | Name, address, phone and hours match everywhere |
| 22–30 | Ask the same questions, same wording, in new threads. Compare with day 1 | Each difference is logged next to what you changed |
Day 30 is a checkpoint, not a finish line. Perplexity doesn't publish how soon it re-reads a page; its model schedules visits by "the importance and likely update frequency" of each URL (Perplexity Research). Repeat the check every month.
How do you know it's working?
Perplexity gives site owners no report like Google's Search Console. Watch three things instead:
- Your question sheet. Same questions every month; count citations and wrong facts. The 20-question ChatGPT test works the same way here.
- Analytics. Google Analytics 4 has an AI Assistant channel, but Google's examples (ChatGPT, Gemini, Deepseek, Copilot, Grok) don't name Perplexity (Analytics Help: Default channel group). Also filter the session source for "perplexity". Perplexity's website uses a "strict-origin" setting, which passes its address only to https sites (MDN: Referrer-Policy), so a site on plain http sees those visits with no source.
- Your callers. Ask how they found you. Someone who read a Perplexity answer may phone without clicking.
Example
An HVAC company behind Bot Fight Mode
Its robots.txt allows every bot, yet the logs show no successful PerplexityBot visits. Cloudflare's Bot Fight Mode is on, and Perplexity's docs say a custom rule can't get around it. The owner turns it off and adds a rule that allows Perplexity's two IP lists.
A bookkeeper in Calgary
Asked for "bookkeeper for contractors in Calgary", Perplexity cites a directory and two competitors. Her services page opens with "Tailored financial solutions." She rewrites it: "Bookkeeping and GST filing for Calgary contractors, from CA$300 a month." By day 30 her own page isn't cited yet, but Perplexity now gives her correct hours from her updated Yelp page. She keeps the monthly check.
A dental clinic with the wrong hours
Perplexity says the clinic is closed Saturdays. It started Saturday hours a year ago. The old hours sit on two directory listings and an outdated contact page. The clinic fixes all three and asks again the next month.
FAQ
Is Perplexity SEO different from Google SEO?
The page work is the same. The difference is access: ranking on Google doesn't prove Perplexity's own crawler can read your site.
Does Perplexity train AI on my website?
Not for pre-training, Perplexity says: it "does not build foundation models, so your content will not be used for AI model pre-training" (Perplexity Help Center: How does Perplexity follow robots.txt?).
Does schema markup help in Perplexity?
Perplexity's docs and help center don't mention schema markup. Its research team does describe reading visible lists and tables, so put key facts there.
Can I keep Perplexity out completely?
Not with robots.txt alone. It stops PerplexityBot's full-text indexing, but Perplexity may keep your domain, headline and a short summary, and Perplexity-User generally ignores it.
Different for you?
- Your customers mostly use ChatGPT: how to get ChatGPT to recommend your business.
- You want one checklist that also covers Gemini, Claude and Copilot: how to rank in Perplexity, Gemini and Claude.
- You're deciding which AI bots to block: should you block AI crawlers?
- Your listings disagree with each other: local citations and directories.
- You'd rather have it done and tracked: SEO & GEO Agent.