An AI visibility tool checks whether AI engines mention, rank and recommend your brand when buyers ask about your category. MentionRadar runs your tracked prompts every week in ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews and turns the answers into one visibility score.
Overall score
64
+7 vs last month
Every number in an AI visibility tool starts as a single answer. This is what MentionRadar stores each time a tracked prompt runs.
You choose the questions your buyers ask, such as "best CRM for a ten person sales team" or "alternatives to a large enterprise CRM". MentionRadar sends each one to every engine on your plan once a week and keeps the full answer text, so you can always read what the engine actually said.
From that text it reads the verdict: is your brand in the answer, where in the list it sits, whether the engine recommends it and which competitors share the answer. On Growth and higher plans it also stores the pages the engine cited, and marks the ones on your own domain.
Because the prompt set is fixed, this week's answer can be compared with last week's. That comparison is the whole point of tracking, and it is what a one-off manual check in a chat window cannot give you.
“What is the best CRM for a ten person sales team?”
For a small sales team, Lumen CRM is a strong pick thanks to its simple pipeline view and fast setup. Brightpath is a good alternative if you need marketing automation in the same tool, and Fieldnote suits teams that live in email.
AI visibility score
64
out of 100
Position history on its own is the job of an AI rank tracker. The visibility score combines all four signals, which is why it moves when a competitor pushes you out of an answer even if your list position looked stable. If you are comparing trackers before you buy, we lay out engines and add-on costs side by side in Otterly AI alternatives and pricing and Peec AI alternatives.
| Verdict for one answer | Points |
|---|---|
| Recommended (top three or called the best pick) | 100 |
| Named fourth in the list | 55 |
| Named fifth or lower | 40 |
| Not named | 0 |
Each answer gets points. A recommendation is worth 100. A mention without a recommendation at position n is worth 100 minus 15 for every place below the first, with a floor of 40, so being named at all always beats being absent. An answer that does not name you is worth 0.
The score is the average of those points over the latest run of every prompt and engine pair from the last eight days. With 15 prompts on three engines that is 45 answers; with 40 prompts on five engines it is 200. Deleted prompts drop out, and an older run never counts twice.
Here is the calculation for a small set. Three prompts on three engines give nine answers: four recommendations (400), one fourth place (55), one seventh place (40) and three answers without your brand (0). The total is 495 points over nine answers, a score of 55. The per engine bars use the same rule restricted to one engine, which is how you spot that Claude is the weak spot while ChatGPT is fine.
The full method, including share of voice and sentiment, is on the methodology page.
AI visibility score in MentionRadar
AI share of voice in MentionRadar
Prompts naming no brand of yours
Answers citing your domain
You moved up in answers where you were already present. Good, but the prompts you are missing from are unchanged. Check the matrix for the rows still marked not named.
You appear in more answers but low in the list. The engines know you, they do not yet prefer you. Comparison pages and reviews that rank you are usually the missing piece.
Engines read different sources. Open the cited pages for the lagging engine and compare them with the ones the strong engines use. The share of voice tool view shows the same split against each competitor.
For every tracked prompt and every engine it records four things: whether your brand was mentioned, its position in the answer, whether the engine recommended it, and which competitors were named. On Growth and higher plans it also stores the URLs the engine cited. Those records roll up into an AI visibility score per engine and an overall score.
AI engines are not fixed lists. Models are updated, engines with web search read new pages, and the same question can return a slightly different answer on different days. That is why MentionRadar tracks a fixed set of prompts on a weekly schedule and charts the trend. One answer is an anecdote, twelve weeks of answers is a pattern.
Most single brands get a clear picture from 15 to 40 prompts. Starter tracks 15 prompts weekly in ChatGPT, Perplexity and Google AI Overviews. Growth tracks 40 prompts in all five engines, and Agency tracks 150 prompts spread across client workspaces.
Position is one of the signals, so there is overlap. An AI visibility tool looks wider: whether you are named at all, whether you are recommended, who else is named and how that compares with your share of voice on the web. If position history is your main question, see our AI rank tracker page.
Both, on every plan. Web mentions from news sites, blogs, forums, Reddit, Hacker News, review sites, YouTube and X arrive every 15 minutes with sentiment and reach. Share of voice is shown for the web and for AI answers side by side.
Yes. Scan your brand from the homepage and you see real web mentions and an AI engine check for your brand before creating an account. Weekly prompt tracking and alerts start with a paid plan, from $24 a month billed yearly.
Scan your brand from the homepage. You see real web mentions and an AI engine answer for your brand in about a minute, before you create an account.