Measuring your AI share of voice
a free monthly prompt panel that shows which assistants name your product, which name competitors, and which sources they read.
last updated 4 November 2026
you measure AI share of voice by asking. once a month, run the questions your buyers actually ask through the assistants they actually use, log which products get named and which sources get cited, and track the change. it is manual, it is free, and at the stage most products are at, it tells you more than a paid GEO dashboard will.
The monthly prompt panel
the panel is a fixed set of questions, run the same way every month, so the only variable is the answers.
- collect 20 to 30 real buyer questions. “what app should i use for x”, “best x for y”, “is there an alternative to z”. pull them from support messages, sales calls, and the Reddit threads where your users already ask them. do not write the questions you wish buyers asked.
- run every question through ChatGPT and Perplexity at minimum, adding Claude or Gemini if your audience uses them. use fresh sessions with no history and memory off, because a session that knows you is a poll of yourself.
- log the products named, per prompt, in a plain sheet: date, assistant, question, products in the order they appeared. order matters, first-named is the real recommendation.
- log the cited sources on the assistants that show them. the domains that keep appearing are your market’s citation map, and that map is your to-do list for earning third-party presence.
- score the month. named-in share, the percent of prompts where you appear at all, is your AI share of voice. track it next to first-named share, and the same two numbers for your top competitor.
the first run takes an afternoon. every run after takes about an hour, and the sheet becomes the only honest record you have of whether any of this work is moving.
panel hygiene is what makes the numbers comparable. keep the wording of every question frozen, because a rephrased question is a different question and its answers do not belong on the same trend line. run the panel the same week each month, from the same account state, and when you discover a new question worth tracking, append it as a new row instead of editing an old one. the discipline feels fussy until the first time a number moves and you need to know whether the market changed or you did.
Reading the log
judge the slope, not the month. a 25-prompt panel is a small sample, and single-month wiggles are mostly noise, so what you are looking for is direction across three or four runs. three things in the log are worth acting on. a competitor entering answers they were absent from is an early warning long before it shows in revenue. a source domain gaining share tells you where next month’s outreach goes. and the first time one of your own pages appears as a citation, look hard at which page and which paragraph got quoted, because that is direct feedback into how you write the rest.
per-assistant differences in the log are not noise either. they usually mirror each assistant’s source diet, which is covered in how AI assistants actually pick products, and they tell you which lever to pull for which assistant.
keep an eye on the prompts where no product gets a confident recommendation at all, the answers that hedge with “it depends” and name nobody. those are open questions in your market, and they are the cheapest share of voice available, because you are competing with silence instead of an incumbent.
Free first, tools later
do not start with a paid GEO tracker. the tools automate exactly what the manual panel does, at scale you do not need yet, and they skip the part where running the prompts yourself teaches you how your market actually asks and who actually gets recommended. that education is the point at this stage.
pay when the panel has proven there is movement worth tracking daily, when you need hundreds of prompts or several markets, or when someone other than you needs the dashboard. until then the sheet wins, because it is free, it is yours, and you understand every number in it.
measurement closes the loop on everything else in the AI search playbook. the breakdown call is free. Let’s talk Growth.
this page is part of the AI search & GEO playbook.
More in this playbook
How AI assistants actually pick products
ChatGPT recommends a tiny fraction of the products it knows. what separates the named from the ignored, and why it differs per assistant.
AI search & GEOWriting answer-first content that gets quoted
where AI assistants pull quotes from on a page, and how to write the 40 to 60 word answers that get cited instead of skimmed past.
AI search & GEOReddit's role in AI answers, honestly
Reddit shows up in roughly 40% of AI citations and the share can drop overnight. how to use the channel without building on sand.
AI search & GEOThe technical layer: what matters and what is theater
static HTML, fast pages and visible answers move AI citations. schema and llms.txt mostly do not. what to build and what to skip.
AI search & GEOListicles and comparison pages AI loves to cite
why lists and comparison tables get cited so often, how to write an honest one, and how to get into other people's.
AI search & GEOAI directories: six good ones beat sixty
a handful of real AI directories beat mass submission. what a listing actually does for AI visibility and how to submit cleanly.
AI search & GEOThe honest GEO timeline
expect 8 to 16 weeks before AI assistants start naming you. what compounds, what spikes, and why a young domain still has a shot.