Method

How to measure your brand’s visibility in AI answers

To measure your brand’s visibility in AI-generated answers, run a repeatable five-step method: (1) write the buyer questions people ask an assistant before they have heard of you; (2) run each across ChatGPT, Perplexity and Gemini; (3) record whether you were named, in what position, and who was named instead; (4) note which sources each engine cited; (5) track the mention rate over time so you can tell a real change from run-to-run noise. To find which prompts make AI ignore your brand, that is step 3: the questions where you are never named, worst first, are your gap list. You can do all of this by hand - tools only automate the repetition.

The five-step method

1. 1

   Write the buyer questions, not your brand name

   List 10 to 20 questions a buyer would ask an assistant before they have heard of you - "best X for Y", "what tool does Z", "how do I…". The questions worth measuring are the ones where your name would not appear because the buyer does not know it yet. Searching your own brand name proves nothing: of course you show up for that.

2. 2

   Run each question across the engines that matter

   Paste each question into ChatGPT, Perplexity and Gemini as a real user would, in a fresh session with no personalisation carried over. These three answer differently because they retrieve differently, so a single engine is a single data point, not a measurement. Do the whole set on each.

3. 3

   Record whether you were named, where, and who beat you

   For each answer, write down three things: were you named at all, in what position (first line or buried at the end), and which competitors were named instead. That last column is the one people skip and it is the most useful - the names that keep recurring are the brands AI treats as the category, and the questions where you never appear are your gap list.

4. 4

   Note the sources each engine cited

   Perplexity shows its citations inline; ChatGPT and Gemini reveal the pages they read when they search. Log the URLs. The pages that keep reappearing across your questions are the pages AI reads in your space. Getting named on those, or out-answering them, is the concrete target - far more actionable than "be more visible".

5. 5

   Track the mention rate over time, not once

   Turn it into one number per pass: named in how many of your questions. Run the set again in a few weeks. A brand named in 2 of 10 answers in January and 5 of 10 in March has evidence something worked; a single run has an opinion. AI answers vary run to run, so the trend line is the measurement and the single check is noise.

How to find the prompts that make AI ignore your brand

This is the half of step 3 people actually came for. When you record every answer, the questions where your brand never appears are the prompts where AI ignores you - and the competitors filling that space are who it names instead. Do not treat them as one flat list. Sort them by buying intent: a prompt like “best tool for X” is worth ten times an idle “what is X”, because the person asking the first one is about to choose. Your fix list is the high-intent prompts where you are absent and a competitor you could honestly beat is named. That is a short, specific, actionable list - not “do more AEO”.

Why measure it more than once

The single most common mistake is running the check once, screenshotting a bad answer, and treating it as the truth. AI answers vary between runs on the same question - Perplexity’s citations in particular shift pass to pass - so one check is a data point and only the trend is a measurement. Pick a cadence you will actually keep (weekly is plenty; more often just measures noise) and compare mention rate against the last pass, not against a feeling.

When to reach for a tool

Everything above is doable by hand, and doing it once by hand is the best way to understand what any tool is really doing for you. What you cannot do by hand is keep it up: 15 questions across 3 engines every week, logged and diffed, is where people quietly stop. That is what the AI visibility trackers automate - Rankscale, Otterly.AI, Peec AI, Profound, Scrunch AI, Semrush and Fulcru. Most of them stop at the number. Fulcru also drafts the page that closes the gap you just found and re-measures after you publish, so the loop from measurement to change stays in one place. See the full comparison with prices before you pick one, and the guide to monitoring your brand’s mentions in AI answers for the tool-by-tool detail.