How do I improve my brand's chances of being cited in Perplexity answers? Our client playbook

Of all the AI engines clients ask about, Perplexity is the one where we can most often show cause and effect. Its answers display their sources, so when a page is cited or dropped, you can see it. That makes it a useful place to test what works. Here is the playbook we use, in the order we use it.

A caution before we start. Nobody outside Perplexity knows exactly how it weighs sources, and we do not pretend otherwise. What follows is practice we can test, not a formula.

Step one: check that the bots can reach the site

Perplexity runs PerplexityBot and Perplexity-User. If either is blocked by robots.txt, a firewall rule or a CDN challenge, nothing else matters. We check the rules first, then the logs, because a page that looks open can still return errors to bots.

We use Promptwatch for this. Agent Analytics logs both bots in real time and verifies IPs, so spoofed user agents are filtered out. Error tracking explains failures, and the crawl-to-citation path gives each page a citation rate. When a client's key product page is fetched by Perplexity-User but never cited, we know to edit the page. When it is never fetched, we fix access.

Step two: find the gap

We then pull the prompts that matter for the client and see who gets cited. Promptwatch monitors Perplexity through the real interface, and shows page and domain citations, URL positions and competitor share of voice. The question we ask is simple: for each prompt, which sources does Perplexity trust here, and what do they have that the client does not?

Often the answer is a type of content. A competitor has a methodology page, a benchmark, or a community thread that answers the question directly. The content gap analysis scores coverage and lists the prompts where rivals appear and the client does not.

Step three: publish something only the client can publish

This is where original research comes in. Answer engines have little reason to cite a page that restates what ten others say. They have reason to cite a page with firsthand data or a clearly described method.

What we have clients do is modest and honest:

• Survey their own customers or community, with the method written down.
• Analyze first-party usage data, anonymized, and publish what it shows.
• Run a benchmark themselves and describe the setup.
• Write explainers that include specifics only the company can provide, such as how a process really works internally.

We never invent a statistic, and we never cite a study we cannot link. A made-up number is worse than no number, particularly on an engine that shows its sources.

Step four: earn mentions outside the site

Perplexity's sources are not limited to brand sites. Articles, forums, reviews and video appear too. We use Promptwatch's offsite mentions report (the brand named in third-party pages without a link) and the Reddit, YouTube and LinkedIn citation reports to see which outside sources the engine already uses for the client's prompts.

Then we decide what is worth doing: a contributed article, a data share with a journalist, a presence in a community where the brand is welcome, a partner page. We do not buy mentions. If a human would not trust the source, we do not want it.

Page Tracker helps here. We add the specific URLs we are working on, ours or a competitor's, and follow their citation behavior.

Step five: write for lifting

On every page we want cited, the direct answer comes first under a heading phrased as the question. Facts that change get dates. Comparisons get a table. Authors are named. For existing pages, Content Refresh proposes diff-style changes so we are editing, not rewriting from scratch.

Step six: measure weekly

We track a panel of prompts by topic and persona, run daily, and review weekly. Citation Trends gives first cited, ramp-up days, peak rate and decay half-life per page, which tells us how fast a new piece is picked up and how long it lasts. Some pages ramp in days, some never do, and the pattern differs from page to page.

Note on plans: Explore, the free tier, is ChatGPT only, so Perplexity tracking needs a paid plan, and crawler logs start on Professional ($245 a month).

What we cannot promise

We cannot promise a citation. Perplexity answers vary between runs, and any tracker is a sample. What we can promise is a clear view of where a client stands, a short list of fixes with evidence behind them, and an honest report on what moved. If you want to run this yourself, Promptwatch is where we do the measuring.