Home / Blog / Guide

Guide

Perplexity SEO: how to rank in AI search

20 June 2026 · Zeb Choudhry

Perplexity AI processes hundreds of millions of queries per month and is growing faster than any other AI search product outside of Google. Unlike ChatGPT, which primarily uses a training-data knowledge cut-off, Perplexity runs live web searches and cites sources in real time. That makes it the AI search product most directly comparable to Google — and most directly susceptible to SEO-style optimisation.

This guide explains how Perplexity selects sources, what PerplexityBot looks for when it crawls your site, and the specific changes that make your pages more likely to be cited.

How Perplexity works (the short version)

When a user asks Perplexity a question, Perplexity runs a web search, retrieves several pages, extracts the most relevant passages, and synthesises a response. It cites the pages it pulled from in a numbered source panel on the right of the answer.

This means Perplexity is doing two things your traditional SEO toolkit does not measure: it is retrieving pages (like a search engine) and deciding whether those pages are clear enough to extract a useful passage from (like a language model).

PerplexityBot: what you need to know

Perplexity sends its own crawler, identified as PerplexityBot in your server logs. Check your robots.txt — if you have a broad User-agent: * disallow rule, or if you have explicitly blocked PerplexityBot, Perplexity cannot index or cite your pages.

The crawler's user agent string:

PerplexityBot/1.0 (+https://docs.perplexity.ai/bots)

To verify PerplexityBot is allowed on your site:

1. Visit yourdomain.com/robots.txt
2. Check there is no Disallow: / under User-agent: PerplexityBot or User-agent: *
3. Check your llms.txt (if you have one) does not exclude AI crawlers from key pages

What Perplexity favours when selecting sources

Freshness

Perplexity's live search retrieval strongly weights recency. For queries about current events, products, or market conditions, a page published or substantially updated within the past 90 days is far more likely to be cited than a three-year-old page — even if the older page has more backlinks. Update your key pages with current data and a visible publication/update date.

Direct, extractable answers

Perplexity's model needs to pull a clean passage from your page. The best-cited pages give a direct answer in a short, self-contained paragraph that could stand alone as a quote. Avoid writing answers that only make sense in context of the surrounding paragraphs. Each section should be independently extractable.

Source authority signals

Perplexity gives higher weight to pages from sources that have established authority in a topic area. This is not purely domain authority (as measured by backlinks) — it includes topical depth and citation frequency. If your site has published consistent, substantive content on a topic for 12+ months, Perplexity is more likely to surface it for queries in that topic cluster.

Clear entity signals

When a user asks "which [service] should I use in [city]" or "what is the best [product] for [use case]", Perplexity needs to understand what your business is, where it operates, and who it serves. Schema markup (particularly LocalBusiness or SoftwareApplication) and a clear, factual About page dramatically improve entity clarity.

Reddit and review site presence

Perplexity frequently pulls from Reddit, Trustpilot, and review aggregators when the query implies social proof or peer opinions. This is not something you control directly — but it is a reason to monitor and respond to reviews, and to participate authentically in relevant online communities where your business is discussed.

Perplexity vs Google: how they differ

Practical steps to optimise for Perplexity

1. Allow PerplexityBot in robots.txt — verify there is no block on the crawler
2. Add visible publication and update dates — freshness signals matter significantly
3. Write in extractable paragraphs — each paragraph should answer one question clearly and stand alone
4. Add schema markup — Organisation, LocalBusiness, or SoftwareApplication depending on your type
5. Publish or update content regularly — dormant sites are deprioritised for live-search queries
6. Build your review profile — Trustpilot and Google reviews get pulled directly
7. Participate in relevant Reddit threads — Perplexity regularly cites Reddit for recommendation queries
8. Create an llms.txt file — a machine-readable summary of who you are and which pages matter most

Monitoring your Perplexity performance

Unlike Google Search Console, Perplexity does not offer a dedicated analytics product for publishers yet. The practical approach is:

• Search for your brand name and key service queries directly in Perplexity and note whether you appear in the source panel
• Watch for referral traffic from perplexity.ai in Google Analytics or your analytics tool
• Use Visus citation monitoring to track whether AI answers start including your business after you make the changes above