Start here
Check PerplexityBot access, inspect the sources returned for your buyer questions, and make your pages useful enough to support an answer. Perplexity also has a user-requested fetcher, so blocking one bot does not establish that a page can never appear.
Perplexity’s source links make it possible to investigate an answer instead of stopping at a mention count. Start by asking which sentence a source supports. A page can be cited for a definition without its company being recommended as a product to buy.
Understand the two access paths
Perplexity’s crawler documentation describes PerplexityBot for search discovery and Perplexity-User for user-requested page visits. The latter generally ignores robots.txt because it responds to a user’s request. Review both paths instead of treating one bot as an absolute citation switch.
For public pages you want discoverable, check robots.txt and CDN rules together. Match published IP ranges as well as the user agent when verifying bot traffic. If a page must remain private, require authentication; robots.txt is not an access-control system.
Build a source inventory from actual answers
- 1.Choose a buying question and record its context, such as country, use case and budget.
- 2.Save the full answer, date, visible citations and search mode. Note whether your brand is recommended, merely mentioned or absent.
- 3.Open each relevant citation and record the page type: vendor documentation, review, comparison, discussion or news.
- 4.Repeat the question and group sources by URL and domain. A recurring source is a lead to investigate, not proof of a universal ranking factor.
Worked example, not customer data
For “inventory software for a two-location bike shop,” one answer might cite an integration guide and a retailer comparison. Check whether your site explains stock transfers and point-of-sale compatibility. Those omissions would suggest a specific page update, rather than a vague instruction to publish more content.
Match freshness to the claim
Keep changeable facts current: pricing, availability, integrations and product versions. Preserve useful historical material with its original dates. Do not bump an article’s date without reviewing its contents. A new date is not evidence that a page is better, and this guide does not assume that fresh content always beats an older source.
Explain the conditions behind a specification. A compatibility claim should identify the supported product and version. A price should identify the plan and billing interval. A statistic should link to its original source and describe what was measured.
Earn third-party coverage where it is relevant
If your saved answers repeatedly use a particular review site or industry publication, check whether it contains a current description of your product. Contribute accurate information through normal editorial processes. Participate in communities as a disclosed representative when appropriate; do not manufacture customer accounts or reviews.
Do not assume Reddit is the right destination for every category. Let the observed source mix for your questions guide the investigation. Different questions can need documentation, local listings, editorial comparisons or first-hand experience.
Measure recommendations and visits separately
Save linked citations separately from unlinked brand mentions. Track referred visits in your analytics where attribution is available. Missing referrers and zero-click reading limit what traffic data can tell you; a citation is not a measured visit.
A Google AI-summary click study cannot supply a Perplexity click-through rate. Use your own attribution data for that question, and keep the denominator visible. Ten visits from one cited page do not establish how many people saw it inside answers.
Put the findings to work
August connects visibility tracking with a reviewable next step: inspect the answers and citations, prepare a draft, publish through your own workflow, then check what changed. Copy the content or create a GitHub draft PR; you control what goes live. A follow-up comparison is evidence to examine, not proof that one edit caused a gain.
Get started with August. Compare the current plans: Starter includes 25 tracked prompts across one brand workspace; Pro includes 75 across three, with Google AI Overviews and AI Mode alongside the four core engines. Prompt allowances are shared across the account.
Related reading
The citation tracking comparison explains what evidence a tracker should preserve. The GEO guide covers the broader content workflow.
Questions, answered.
Does PerplexityBot control every possible visit?
No. Perplexity also documents Perplexity-User for user-requested access. See the provider reference for the distinction.
Does a citation mean Perplexity recommends my brand?
Not necessarily. Read the claim attached to the link and record recommendations separately from source citations.
Do I need to publish on Reddit?
Only if it is relevant to your buyers and the questions you are investigating. Inspect actual sources before choosing an off-site activity.
Find out what AI tells your customers.
Book a call and we’ll show you exactly what AI says about you.