Essay · 08

How do I measure whether my brand is being cited by AI assistants like ChatGPT and Perplexity?

13 July 2026 · 6 min read

To measure whether AI assistants cite your brand, run the real questions your buyers ask through ChatGPT, Perplexity, Claude and Google's AI answers on a fixed cadence, and log three things each time: whether you're mentioned, whether you're cited with a link, and which sources show up instead. That gives you a citation rate and, over time, a trend. There is no out-of-the-box dashboard for this the way rank trackers existed for SEO, so you either build a simple logged process or use a tool or service that does it for you. What matters is that you measure the same questions the same way every time, so the number actually means something.

Here's the deal: most people check once, see themselves missing from a single answer, and either panic or shrug. Neither is measurement. Measurement is a repeatable process that turns "am I cited?" into a number you can watch move. Here's how to build one that gives you an honest answer.

One prerequisite before you measure anything: make sure the engines can read your site at all. If your robots.txt or your CDN blocks the AI crawlers, you'll measure zero for a reason that has nothing to do with your content. My AI Visibility Check tells you in thirty seconds whether that's happening.

Why you can't just check once

Ask ChatGPT the same question twice and you can get two different answers, citing different sources. Ask Perplexity, Google and Claude the same question and you'll get four different pictures. So a single screenshot is an anecdote, not a measurement. To get a signal you can trust, you need two things: a fixed set of questions you ask every time, and more than one run per question per engine, so a one-off fluke doesn't read as a trend. Treat it like polling, not a single vote.

Pick your query set the way your buyers actually ask

Measure the questions your buyers actually type, not your own brand name. "Best [category] tool for [use case]", "alternatives to [competitor]", "who can I hire to [do the thing]". These are the moments where a recommendation gets made, and where being cited actually matters. Checking whether an engine knows your brand when you prompt it with your brand tells you nothing useful. Write down five to ten of these buyer-intent questions and lock the list, because the whole value is in asking the same set over time, so this month is comparable to last.

Measure the right thing: mention vs attributed citation

There's a difference between being mentioned and being cited, and it matters. A mention is your name appearing in the answer's prose. A citation is a structured, linked source the engine attributes a claim to. Both have value, but if you count loose prose mentions you'll flatter yourself and lose comparability, because prose is fuzzy and links are not. Count structured citations only. That deliberately makes your number lower, which is the point: an honest baseline you can defend beats a generous one you can't, and a false positive quietly corrupts the very thing you're trying to prove.

Run it across all four engines, because they disagree

Don't measure one engine and assume the rest. Each has its own taste in sources. Perplexity cites most generously and leans hard on YouTube and community threads. Google's AI Overviews draw heavily on community discussion. Claude and ChatGPT are stingier and favour tidy, self-contained explainers. Measure ChatGPT, Perplexity, Claude and Google's AI answers separately, and keep the numbers separate, because "cited in Perplexity but invisible in ChatGPT" is a specific, actionable finding that an average would hide.

Log it on a cadence and watch the trend

One measurement is a baseline; the value is in the trend. Pick a cadence you'll actually keep, monthly is a sane default, and log each run: the date, the query, the engine, whether you were cited, and which sources appeared instead. Keep the log append-only so you never overwrite your own history. Alongside it, watch AI-assistant referral traffic as its own channel in your analytics. It's a partial signal, because a growing share of AI answers produce no click at all, but a rising number of visits arriving from ChatGPT or Perplexity is a second line of evidence that the citations are landing.

What the measurement tells you to do next

The most useful output of measuring isn't your own score, it's the list of sources cited instead of you. When you log which pages, threads and videos the engines quote for your buyer questions, you're looking at your target list: the specific surfaces you need to be present on, and the format the engines prefer for that question. Measurement isn't a vanity dashboard, it's the front end of the fix, because it converts a vague "we should do AI visibility" into a ranked, concrete list of where to earn a mention first. Muck Rack's Generative Pulse study puts earned media at around 84% of AI citations (Muck Rack Generative Pulse, May 2026), so expect most of that target list to be third-party surfaces, not your own pages.

Where I fit

This is the front end of the work I do. I run exactly this measurement, a logged scan of a fixed buyer-question set across the four engines, as the first step of the Citation Engine, because you can't fix a gap you haven't measured. Case Study Zero runs on this site itself, in public, and it's measured this way: it started at cited in zero of seven buyer queries, and I'm publishing what moves that number as it moves.

The boundary: if you just want the number once, everything above is the method, and you can build it in an afternoon. You hire when you'd rather have it run on a cadence and the gaps it finds actually closed. And if your measurement comes back strong already, you don't need me, which is a perfectly good result to get from an afternoon's work.

If you want the wider method, start with how to get cited by AI search engines. If you're weighing whether to hire it out, who to hire covers how to vet anyone, me included. When you want it run for you, the contact form is the channel.

Common questions

← All essays