How to Get Cited by Perplexity: A B2B AEO Playbook.

How to get cited by Perplexity comes down to two gates most B2B sites never think about: can Perplexity's crawler retrieve your page at all, and if it can, does what it finds actually shape the answer it writes. Pass both and you show up in the citation list. Fail either one and no amount of keyword-style content will help.
Perplexity is not Google with a chat interface bolted on, and it is not ChatGPT either. It runs a live retrieval step on most queries, reads a short list of pages in real time, and writes an answer with footnoted sources. That mechanic rewards a different set of fixes than classic SEO, and a slightly different set again from getting cited in ChatGPT. This playbook covers what to check first, in order, the same order SourceRank AI's done-for-you AEO service works through for clients.
How does Perplexity choose which sources to cite?
Perplexity evaluates a short set of candidate pages per query, typically in the range of eight to ten, and cites only three to five of them. Getting into that shortlist requires clearing two separate gates.
The first gate is retrieval selection: does Perplexity's crawler find and fetch your page in time. The second is answer absorption: once fetched, does your page contain a clear enough answer that the model actually uses it rather than a competitor's page that said the same thing more directly. Most B2B sites fail the first gate silently. Their content might be excellent, but if the page never gets retrieved, absorption never gets a chance.
Why does Perplexity behave differently to ChatGPT and Google AI Overviews?
Perplexity leans harder on real-time web retrieval than ChatGPT's default mode, which often reasons from training data unless it explicitly triggers a browsing tool. Google AI Overviews sits closer to traditional search ranking signals layered under a generated summary.
That difference matters for B2B teams choosing where to prioritise effort first. A page that already ranks reasonably in organic search has a head start with Perplexity, because Perplexity's retrieval step draws on similar crawlability and indexation signals. A page with strong brand mentions but weak on-page clarity can still underperform in Perplexity even if it occasionally surfaces in a ChatGPT answer built from older training data.
What technical barriers block Perplexity from citing your site?
The most common blocker is crawler access: if PerplexityBot is blocked by robots.txt, a firewall rule, or a bot-management CDN setting, nothing else in this playbook matters. Check crawler access first, before touching content.
The second most common blocker is render dependency. Perplexity's fetch step has a limited wait window, and content that only appears after client-side JavaScript execution can be missed entirely. A B2B pricing page or comparison page built as a single-page app with no server-rendered content is functionally invisible to this kind of retrieval, no matter how good the copy is once it loads. Anyone working out how to get cited by Perplexity should check these two blockers before touching a single sentence of copy.
What does Perplexity's own documentation say about its crawlers?
Perplexity runs two crawlers with different jobs, and only one of them is governed by robots.txt. Its crawler documentation describes them like this (Perplexity):
| Crawler | What it is for | Follows robots.txt | Used to train AI models |
|---|---|---|---|
| PerplexityBot | Surfacing and linking websites in Perplexity's search results | Yes, and Perplexity recommends allowing it | No, per Perplexity |
| Perplexity-User | Visiting a page when a user's question calls for it, to help answer and link to that page | Generally not, because a user requested the fetch | No, per Perplexity |
Three practical consequences follow. First, robots.txt alone is not enough: Perplexity recommends allowing PerplexityBot's published IP ranges in your firewall, taken from its official JSON endpoints and refreshed automatically, matching both the user agent and the IP address, with allow rules placed above any blocking rules. It also warns that Cloudflare's Bot Fight Mode cannot be skipped with WAF custom rules, so a site can end up blocking PerplexityBot without anyone deciding to.
Second, blocking PerplexityBot to keep your content out of AI training is a misread. Perplexity says the crawler is not used to crawl content for AI foundation models, so blocking it mainly takes you out of the search results it feeds.
Third, PerplexityBot does not run JavaScript. Vercel's study of AI crawler traffic listed it among the major AI crawlers that render none (Vercel, December 2024), which is the evidence behind the render-dependency step in the playbook below. Google's Gemini is the exception among the big engines, because it uses Googlebot's rendering, as our Gemini citation playbook explains.
How is content structured to earn a Perplexity citation?
Perplexity extracts passages rather than whole pages, so the direct answer to a likely question needs to sit in the first sentence or two of the relevant section, not buried after three paragraphs of scene-setting. Perplexity's answer-writing step also shows a clear preference for HTML tables when the query is comparative, and for genuine ordered lists when the query is a how-to.
| Engine | Primary retrieval style | Favours in content | Ideal structure |
|---|---|---|---|
| Perplexity | Live web fetch, most queries | Direct passages, tables, ordered steps | Answer-first paragraphs, one claim per sentence |
| ChatGPT | Training data plus optional browsing | Consistent third-party corroboration | Entity clarity, repeated brand-fact pairing |
| Google AI Overviews | Search index plus generation layer | Established organic ranking signals | Traditional on-page SEO plus schema |
What is the playbook for getting cited by Perplexity?
Work through these five steps in order. Skipping ahead to content work before fixing access wastes the content.
- 1. Confirm crawler access. Check robots.txt, CDN bot rules, and server logs for PerplexityBot requests and response codes. Fix any block before doing anything else.
- 2. Remove render dependency on money pages. Server-render or pre-render the pages you most want cited: services, pricing, comparison, and FAQ content.
- 3. Rewrite for passage extraction. Put the direct answer in the first sentence of every section. Convert comparative claims into a genuine markdown or HTML table.
- 4. Build entity and schema clarity. Add Organization, FAQPage, and Product or Service schema so the page's claims are machine-readable, not just human-readable.
- 5. Earn third-party corroboration. Perplexity weighs whether other indexed pages say the same thing about you. Directory listings, review platforms, and press mentions all feed this signal, which is exactly what SourceRank AI's citation building work targets alongside on-site structure.
How long does it take to get cited by Perplexity?
Once a page is crawlable and correctly structured, it can appear in a Perplexity citation within days of being indexed, but a single citation is not the goal. Consistent citation across the questions your buyers actually ask usually takes two to three months of combined on-site fixes and off-site authority building, which lines up with how our full AEO process is sequenced.
Do not assume your Perplexity performance matches your ChatGPT or Google performance. The engines disagree sharply about what to cite: in SourceRank AI's September 2026 runs, ChatGPT cited a community site in 0 of its 923 answers while Google AI Overviews did in 142 of 875 (what AI engines cite by industry), so measure Perplexity on its own prompt set. The free AI visibility score checks whether your site is structurally ready to be cited first.
How do you track Perplexity citations?
Track Perplexity citations from three angles, because each one sees something the other two miss. Perplexity searches the web in real time for each question and attaches numbered citations linking to the original sources (Perplexity Help Center), so the raw material is visible on every answer; the work is in recording it consistently.
- 1. A fixed question set, re-run on a schedule. Write 20 to 50 questions your buyers actually ask, run them in Perplexity every two to four weeks, and log whether your company is named, which of your URLs appear among the numbered citations, and which companies are named instead. Keep the wording identical between runs, because a reworded question makes the comparison meaningless.
- 2. Server logs for Perplexity-User. Perplexity-User only visits a page when a user's question calls for it, so its hits are a direct list of the pages Perplexity pulls in to answer real questions, including questions you never thought to test. PerplexityBot hits show crawling, not use. Check both against Perplexity's published IP ranges, since a user agent string on its own can be faked.
- 3. Referral visits in analytics. Filter GA4 for sessions from perplexity.ai to see which cited pages people actually click through to, and what they do after they land.
A log with five columns is enough to start:
| Column | Example | Why it matters |
|---|---|---|
| Question | "best ERP consultants for mid-market manufacturers" | Must stay word for word identical between runs |
| Date | 10 Nov 2026 | Perplexity's answers change between runs |
| Named in the answer | Yes or no | The outcome a buyer actually sees |
| Your URLs in the citations | /pricing/ or none | Shows which page earned the citation |
| Named instead | Competitor names | Your real competition for that question |
Read the log for movement across several runs, not for any single answer. One run where you appear proves little; the same question citing you three runs in a row is a pattern worth acting on. Our guide to how AI visibility is measured covers turning readings like these into a trended number.
Do B2B brands need a different strategy for Perplexity than for ChatGPT?
Not a different strategy so much as a different order of operations. The underlying assets, clear entity data, schema, third-party citations, and answer-first content, serve every engine. What changes is which fix pays off first: for Perplexity it is almost always crawler access and page structure, while for ChatGPT it is more often consistent third-party corroboration of your core claims. Our industry pages break this down further by sector, since a manufacturing buyer's fan-out questions differ from a professional services buyer's.
Frequently asked questions
What is Perplexity AEO?
Perplexity AEO is the set of technical and content changes that make a page retrievable and citable by Perplexity's answer engine specifically, as distinct from general SEO or ChatGPT-focused optimisation.
Does Perplexity use Google's search index?
No. Perplexity runs its own crawler and retrieval pipeline rather than relying on Google's index, which is why a page can rank well in Google and still be missed by Perplexity if it blocks Perplexity's crawler or depends on client-side rendering.
How is getting cited by Perplexity different from getting cited by ChatGPT?
Perplexity leans more heavily on live retrieval for most queries, so crawlability and page structure matter more immediately, while ChatGPT's default mode often draws on training data and rewards consistent third-party corroboration of a claim over time.
Does Perplexity respect robots.txt?
PerplexityBot, the crawler that surfaces and links sites in Perplexity's search results, follows robots.txt, and Perplexity recommends allowing it. Perplexity-User, which fetches a page when a user's question calls for it, generally ignores robots.txt because a user requested the fetch.
Is PerplexityBot used to train AI models?
No. Perplexity's crawler documentation says PerplexityBot is not used to crawl content for AI foundation models; it exists to surface and link websites in Perplexity's search results.
Can you pay Perplexity to be cited?
No. Perplexity's core citations come from its organic retrieval and answer-generation process, not a paid placement. Any service claiming guaranteed paid citations is not describing how the product works.
Does Perplexity crawl JavaScript-rendered pages?
Perplexity's fetch step has a limited wait window and can miss content that only renders after client-side JavaScript executes, so server-rendering or pre-rendering your most important pages meaningfully improves citation odds.
How many sources does Perplexity typically cite per answer?
Perplexity generally evaluates a shortlist of around eight to ten candidate pages per query and cites roughly three to five of them in the final answer.
Should a B2B company prioritise Perplexity over ChatGPT?
Prioritise whichever engine your buyers actually use to research your category. Checking your current citation rate by engine is a faster way to decide than guessing, since the gap is often uneven across engines for the same brand.
How do you measure your Perplexity citation rate specifically?
Track a fixed set of prompts your buyers would realistically ask, run them against Perplexity on a schedule, and record whether and how your brand is cited. Add Perplexity-User hits from your server logs, which show the pages Perplexity fetches to answer real questions, and referral visits from perplexity.ai. SourceRank AI's measurement approach and the combined ChatGPT and Perplexity citation guide both walk through this in more detail. If tracking this by hand across engines and prompts sounds like a lot of manual work, that repeatable measurement is also exactly what our managed AEO service automates.