Guide

How does Perplexity choose which businesses to recommend?

Updated · 10 min read · By the SearchCrow team

Perplexity recommends businesses by searching the web live when a customer asks, pulling information from pages it treats as authoritative, and writing an answer with numbered citations that link to those pages. It runs its own index of "hundreds of billions of webpages" and scores small sections of each page against the question, so a business gets recommended when its own site and the sites that mention it answer the question clearly and can be crawled. Perplexity has not published a ranking formula.

Key takeaways

  • Every Perplexity answer shows its sources inline, so you can see exactly which websites it trusts for your trade.
  • Perplexity has two crawlers. PerplexityBot builds its search index; Perplexity-User fetches pages when a user asks something. Blocking PerplexityBot keeps you out of its results.
  • Perplexity breaks pages into small units and scores each one, so the paragraph that answers a question matters more than the page as a whole.
  • Perplexity's app shows places on a map with photos, hours and reviews, but it has not said which review sources it uses.
  • The work that gets you cited by Perplexity (crawlable pages, clear facts, consistent listings, reviews) is the same work that helps every other AI assistant.

How does Perplexity answer a question?

Perplexity searches the internet at the moment you ask, then summarises what it finds with a citation on every claim.

Perplexity calls itself an answer engine: rather than returning a list of links, it works by "identifying trusted sources" and combining what it finds into one response (Perplexity Help Center, 2026). Its help centre says it will "search the internet in real-time", "gathering information from authoritative sources like articles, websites, and journals", and that "each answer includes numbered citations linking to the original sources" (Perplexity Help Center, 2026).

Two things follow from this for a business:

  • There is no stored "answer" about you. Perplexity looks you up each time. Change your website or your listings and the next answer can change too.
  • The sources are public. Each citation is a website Perplexity chose for that question. If you are not named, the cited sites tell you where it looked instead.

Perplexity also has deeper modes. Its help centre describes Pro Search, which researches more sources and is applied automatically if your plan includes it, and Deep Research, which runs dozens of searches and reads hundreds of sources (Perplexity Help Center, 2026). The more sources a mode reads, the more chances a well-documented business has to be found.

Where does Perplexity get its search results?

From its own web index, which it also sells to developers as the Perplexity Search API.

When Perplexity launched that API on 25 September 2025, it described "an index covering hundreds of billions of webpages" and said the API runs on "the same global-scale infrastructure that powers Perplexity's public answer engine" (Perplexity, 2025). The same post gives three details that matter if you want to be found:

  1. It is updated constantly. Perplexity says it processes "tens of thousands of index update requests" every second to keep results current. New pages and changes can show up quickly.
  2. It parses messy pages with AI. An AI-powered content understanding module generates parsing logic for pages that are hard to read, and improves itself using signals from real user queries. Clean, text-based pages give it less to untangle.
  3. It scores parts of pages, not whole pages. Documents are split into fine-grained units that are "individually surfaced and scored against the original query parameters". In plain terms, Perplexity is looking for the passage that answers the question, and a page can be cited for one good paragraph.

Perplexity's developer documentation also shows that search can be personalised by location: the Web Search tool accepts a user_location with country, region, city and coordinates, and "city and region significantly improve location accuracy" (Perplexity, 2026). That is the developer API rather than the consumer app, but it confirms the system is built to return location-specific results.

What are PerplexityBot and Perplexity-User, and what happens if I block them?

Perplexity has two crawlers with different jobs, and only one of them reliably obeys robots.txt.

Both are documented on Perplexity's crawlers page (Perplexity, 2026):

CrawlerWhat it doesFollows robots.txt?If you block it
PerplexityBot"Surfaces and links websites in Perplexity search results". Perplexity says it is not used to crawl content for AI foundation modelsYesYour site is not added to the index, so Perplexity is far less likely to find and cite you
Perplexity-UserVisits a page when a user's question needs it, to help answer and link to that page. Not used for crawling or training"Generally ignores robots.txt rules", because a user requested the pageLittle effect on robots.txt alone; Perplexity says this setting controls which sites user requests can reach

Three practical points from the same page:

  • Each setting is independent. Allowing one crawler does not allow the other.
  • Changes take time. Robots.txt changes "may take up to 24 hours to take effect".
  • Your firewall matters as much as robots.txt. Perplexity publishes IP ranges for each crawler (perplexitybot.json and perplexity-user.json on perplexity.ai) and asks sites to allow them. It warns that Cloudflare's Bot Fight Mode cannot be skipped by custom rules and may still challenge its bots. A CDN with "AI bot blocking" turned on can make you invisible to Perplexity even with a perfect robots.txt.

The robots.txt check takes two minutes: open yourdomain.co.uk/robots.txt and look for a Disallow: / line under User-agent: PerplexityBot. Our AI visibility checklist has a copy-paste example that allows the search crawlers for every major assistant.

What kinds of pages does Perplexity cite?

Perplexity documents the shape of what it prefers (authoritative, current, clearly parsed, answering the question in a specific passage) but not a list of favoured websites.

What is documented:

  • Authoritative sources such as "articles, websites, and journals" (Perplexity Help Center, 2026).
  • Passages, not pages. Because the index scores fine-grained units of a page, the paragraph that directly answers "how much does a boiler service cost in Leeds?" (an illustrative example) is what gets surfaced (Perplexity, 2025).
  • Established publications. Perplexity's Publishers' Program, announced on 30 July 2024 with TIME, Der Spiegel, Fortune, Entrepreneur, The Texas Tribune and WordPress.com, shares revenue when an answer references a partner's content (Perplexity, 2024). Perplexity has not said partners rank higher, but its answers plainly lean on recognised publications.

What practitioners observe (not confirmed by Perplexity)

People who track Perplexity citations commonly report that it cites review platforms, directories, "best of" articles from local and trade publications, and forum threads, with the mix changing by question. Perplexity has not confirmed any of this and answers differ from run to run. Treat these as places to look, then confirm with your own citations from the five-question check below.

Does Perplexity use reviews and business listings?

Perplexity's app now shows places with photos, hours and reviews, but it has not named the sources behind them.

Over 2025 Perplexity added a run of local features, documented in its changelog:

  • Places on iOS, where users can "browse photos, check hours, and read reviews—all in one experience", and inline hotel cards with "photos, key details, and quick actions" (Perplexity, 2025).
  • Restaurant booking: "You can now book restaurants directly on Perplexity through our OpenTable integration", with search by cuisine, vibe or location (Perplexity, 2025).
  • Maps for places search: "Looking for a restaurant, hotel, or travel destination? You'll now see them on a map across Web and iOS" (Perplexity, 2025).

None of these updates says which provider supplies the photos, hours and reviews, so we don't name one. The safe reading is that a restaurant, hotel or venue with missing hours, no photos or few reviews on the big platforms will look thin wherever that data comes from. For other trades, Perplexity's answers still rely on ordinary pages it can crawl: your site, directories and review sites. Our Google Business Profile and reviews guide covers getting those right.

Make your site crawlable, answer customer questions in plain text, and be present on the listings and sites Perplexity already cites for your trade.

  1. Allow PerplexityBot and check your firewall. Remove any Disallow for PerplexityBot in robots.txt, and check your host or CDN for "AI bot" or "bot fight" settings that block Perplexity's published IP ranges (Perplexity, 2026). Allow a day for changes to apply.
  2. Put the facts in text, not images or scripts. Services, prices or "from" prices, areas covered, opening hours, qualifications and contact details should be readable in your page source. Perplexity's parser is built to cope with messy pages, but there is no reason to make it work for it.
  3. Answer the question in the first paragraph. Perplexity scores passages. For each service page, open with two or three sentences that say what you do, where, for whom and roughly what it costs, as if replying to a customer who asked. Put detail below, not above.
  4. One page per real question. "Emergency plumber in Harrogate", "boiler service cost", "do you cover Knaresborough?" (illustrative examples) each deserve a clear page or section with a heading that matches the question. Thin pages stuffed with every town name do the opposite.
  5. Keep your name, address and phone number identical everywhere. Website, Google Business Profile, Bing Places, Apple Maps and the directories for your trade. Perplexity has not documented a consistency check, but an assistant that finds two phone numbers has less reason to trust either.
  6. Get listed and reviewed where Perplexity looks. Run the five-question check below and note the cited sites. If a directory, review platform or local publication keeps appearing and you are not on it, that is your next job. Ask every customer for a review, never only the happy ones.
  7. Show you're real and qualified. An about page with named people, trading history, accreditations with registration numbers and links to the issuing bodies gives Perplexity a trustworthy page to cite when a customer asks "is [business] any good?".
  8. Keep it current. Perplexity re-searches every time, so stale hours or prices will be repeated to customers until you fix them.

Our guide to how AI assistants choose businesses covers the other assistants, and the AI search optimisation guide goes deeper on writing pages that get cited.

How do I check whether Perplexity recommends my business?

Ask it five questions your customers ask and read the citations, not just the answer.

  1. Write five questions in the words a customer would use. For a dentist in Bristol, illustrative examples might be "best private dentist in Bristol", "emergency dentist Bristol today", "how much is a dental implant in Bristol?", "is [your practice] any good?" and "[your practice] vs [competitor]".
  2. Ask each one in Perplexity (the free version is fine) and note three things: whether you are named, whether what it says about you is accurate, and every website in the numbered citations.
  3. Count the citations by site. The sites that appear across several answers are the ones Perplexity trusts for your trade in your town. Being absent from them is usually the biggest gap.
  4. Repeat on another day. Answers vary between runs, so one screenshot proves little. Three runs of five questions is a fair picture.
  5. Fix the biggest gap first, then check again a month later.

That is manageable by hand for one assistant. It gets tedious once you add ChatGPT, Gemini, Claude and Google's AI features, and competitors too.

Why does Perplexity matter if my customers mostly use Google?

Because it is the clearest window into what AI assistants think of your business, and the fixes carry over.

  • It shows its working. Perplexity's numbered citations on every answer make it the easiest assistant to audit. If Perplexity cannot find a reason to recommend you, the others probably struggle as well.
  • It reacts fast. A continuously updated index (Perplexity, 2025) means you can see the effect of a fix within days.
  • It is spreading beyond the chat box. Perplexity launched the Comet browser on 9 July 2025 with its assistant built in (Perplexity, 2025), and its index now powers other companies' products through the Search API. Being findable by Perplexity means being findable in more places than perplexity.ai.
  • Same work, every assistant. Crawlable pages, clear facts, consistent listings and genuine reviews are what every assistant reads.

For the Google side, see our guide to Google AI Overviews and AI Mode for small businesses; for ChatGPT, see how to get recommended by ChatGPT.

How SearchCrow helps

Perplexity is one of the two assistants in SearchCrow's free check: enter your website and it asks ChatGPT and Perplexity five questions a customer might ask, checks your site the way AI crawlers see it, and shows whether you were named, who was named instead, and the top three fixes. The paid plans repeat the check on a schedule, add the other assistants, and list the sites Perplexity cites that you are missing from. Everything in this guide works without it; SearchCrow just saves you the monthly hour of asking.

FAQ

How does Perplexity decide which businesses to recommend? Perplexity searches the web in real time when a question is asked, gathers information from sources it treats as authoritative, and writes an answer with numbered citations. It runs its own index of hundreds of billions of pages and scores small sections of each page against the question. It has not published a ranking formula, so the practical route is to be crawlable, clear and well documented on other sites.

Should I block PerplexityBot in robots.txt? Not if you want to appear in Perplexity answers. Perplexity says PerplexityBot surfaces and links websites in its search results and is not used to collect content for training AI models. Perplexity-User, the second crawler, fetches pages when a user asks a question and generally ignores robots.txt. Perplexity says robots.txt changes can take up to 24 hours to take effect.

Does Perplexity use reviews and business listings? Partly documented. Perplexity's 2025 product updates added place results on a map for restaurants, hotels and destinations, an iOS places experience where users can browse photos, check hours and read reviews, and an OpenTable integration for restaurant bookings. Perplexity has not said which review sources feed these features, so keep your reviews and listings accurate on the major platforms anyway.

How can I check whether Perplexity recommends my business? Open Perplexity, ask five questions a customer might ask, such as the best provider of your service in your town, and read each answer's numbered citations. Note whether you are named, whether the details are right, and which websites are cited. Repeat the questions another day, because answers vary between runs. The cited sites are where to focus your listings and reviews.

Why does Perplexity matter if my customers use Google? Perplexity is a smaller audience than Google, but it is a clear test of whether the web says enough about you for an AI assistant to recommend you with confidence. Because every answer shows its sources, it tells you exactly which sites it trusts for your trade. The same crawlable pages, listings and reviews also help ChatGPT, Gemini, Claude and Google's AI features.

Sources

Free check

Does ChatGPT recommend your business?

Free, in under a minute. You’ll see whether AI assistants mention you, who they recommend instead, and your top 3 fixes.

We’ll read your website for your business details, so there’s less to type.

Free, no card needed. We ask ChatGPT and Perplexity 5 questions your customers might ask, and check your website. We use your email to send you your results and a few short follow-ups about them, which you can stop with one click; see our privacy notice.