Your buyers are asking assistants which tool to use, and you have no dashboard telling you what they hear. The gap is not unfixable, but closing it requires building the measurement yourself, because nobody is going to hand you the data.
Step one: build the question set
Everything depends on asking the right questions, and the right questions are the ones a buyer would actually type — not your keyword list.
Build twenty to thirty covering five shapes: category discovery ("best tools for [job]"), segmented ("best [category] for a 10-person agency"), competitor alternatives ("[competitor] alternatives"), evaluation ("is [competitor] worth the price"), and problem-first ("how do I stop [specific pain]") where the buyer does not yet know the category exists.
That last shape is consistently the most revealing and the most neglected, because it captures demand before anyone has searched a product term.
Step two: run it consistently
Run every question across ChatGPT, Perplexity, Google AI Overviews, and Claude, monthly, in a fresh session each time with no personalization or memory carried over. Model outputs vary, so run each question two or three times and record the distribution rather than a single result.
Capture four fields per run: were you named, which competitors were named, in what order, and which sources were cited. A spreadsheet is entirely sufficient for this.
Step three: the three metrics that matter
Mention rate. The share of your question set where you are named at all. This is the headline number, and for most companies starting out it is sobering.
Share of voice. Your mentions as a proportion of all vendor mentions across the set. This is the competitive picture, and it is the one executives respond to.
Source inventory. The ranked list of every URL cited across all runs. This is the actionable one. It converts "we need better AI visibility" into a specific list of pages to outrank and threads to join.
Step four: act on the source inventory
Sort the cited sources into three buckets.
Pages you control. If they are cited, make them more quotable — tighter answers, more specific claims. If they should be cited and are not, that is usually a clarity problem rather than an authority one.
Third-party editorial. Roundups, review sites, comparison articles. If you are missing from one that gets cited repeatedly, getting included is a direct, bounded piece of work.
Community threads. Almost always the largest bucket for segmented and evaluation questions. You cannot edit these, and getting into old ones has little effect. The lever is being present for new ones as they appear, which is a monitoring problem rather than a content problem.
Setting realistic expectations
Improvements to your own pages can show up within weeks in browsing-enabled systems. Changes in what a model says without browsing wait on retraining, which is a matter of months.
Community presence compounds and does not decay: a thread you contributed to usefully this quarter continues to be retrieved indefinitely. It is the slowest input and the one with the longest tail, which is a good argument for starting it before you feel ready.
FREQUENTLY ASKED QUESTIONS
Questions about measuring AI visibility
Are there tools that track AI mentions automatically?
A growing number of products poll assistants on a schedule and report mention rates. They save time, but the question set still has to reflect how your buyers actually phrase things, and that part cannot be outsourced.
How often should I measure?
Monthly is enough for most companies. Weekly produces noise, because run-to-run variance in model outputs is large relative to real change.
What is a good mention rate?
There is no universal benchmark, and the absolute number matters less than the trend and your share relative to named competitors on the same question set.
Should I optimize for one assistant?
Optimize for the underlying mechanism rather than one product. Being retrievable, quotable, and genuinely well-discussed improves visibility across all of them, because they share those inputs.
CLOSE THE LARGEST GAP
Community threads are your biggest source bucket. Work them.
MentionSpot monitors Reddit and X for the questions that feed AI answers in your category, scores them for buying intent, and delivers a ranked queue.
Get the 7-day pass