How we measure

What each number means, and why you can rely on it.

AI visibility is how often, and how favorably, AI assistants name your brand when people ask the questions your buyers ask. This page explains how BotSay measures it. It is never behind a form.

Written by the BotSay team. Last reviewed Sep 26, 2026.

To see each number on its own screen, read What the five measures mean and How BotSay measures your numbers in the docs.

  1. The five measures, in one line each
  2. How sure we are of a number
  3. How we know a change is real
  4. Which questions count toward which number
  5. How each engine is asked
  6. Why every week
  7. How sentiment is scored
  8. What is outside our reach

The five measures, in one line each

Mention rate. Of the answers we read, the share that named you.

Share of voice. Of every brand mention in those answers, the share that was you.

Average rank. When you are named, how early. Lower is better.

Your pages cited. Of every link the answers gave, the share that pointed at your sites.

Sentiment. How favorably you are described, from 0 to 100, scored by a model.

How sure we are of a number

Mention rate is counted from a known number of answers, and we print that count beside it. From that count we work out a likely range. If your Mention rate is 34% with a range of 31 to 37, your true rate is almost certainly somewhere in that span. More answers make the span narrower, which is why we read every question on every engine each week. The other four measures, Share of voice, Average rank, Your pages cited and Sentiment, show a figure and its movement, without a range. These are numbers you can stand behind. Our guide How many questions do you need to track AI visibility? shows what each plan gives you.

How we know a change is real

A change in Mention rate from one week to the next is real once it is bigger than chance alone could explain, given how many answers each week is counted from. In this example, last week's 27.8% was likely between 25.1 and 30.7, and this week's 34.0% is likely between 31.1 and 37.0. The two ranges do not overlap, so BotSay calls it what it is: a real change, up 6.2 points.

A steady week is useful too. When nothing moved enough to count, the reading says so and tells you how big a change would have needed to be, so you know where the bar sits for next week.

Mention rate, two readings
The figures as a table
Mention rate, two readings
WeekMention rateLikely range (%)
Last week27.8%25.1 to 30.7
This week34.0%31.1 to 37.0
Up 6.2 points. A real change.

Example, using an invented brand

Which questions count toward which number

Some questions name you already, like "Is [your brand] any good?" AI names you in almost every answer to those, so we report them on their own. Your headline numbers come from questions like "best project tools for agencies", where being named is something you earned.

How each engine is asked

We ask ChatGPT, Perplexity, Gemini, Claude, Google AI Overviews and Google AI Mode through each provider's official developer access or a licensed search-data provider, with web search switched on where the engine offers it, in English. Developer access is the official route AI companies give software for asking questions, rather than typing into the chat app. Each question is asked as the country you choose for it, the United States unless you pick another. Gemini cannot be pointed at a country, so it gives the same answer wherever it is asked from. A question asked in a new country is a new question, so its answers are never mixed with the old country's. Every engine gets the same questions in the same week.

Why every week

A complete weekly reading means every week compares cleanly with the last: same questions, same engines, same method. That is what lets us say with confidence that you moved. A week is compared with the one before only when every engine answered in full, and the reading tells you when one did not, so every comparison you see is like for like.

How sentiment is scored

A language model reads each sentence that names you and scores how favorably it describes you, from 0 to 100. We label the figure as estimated wherever it appears.

What is outside our reach

We measure what the engines answer when asked through each provider's official developer access or a licensed search-data provider. A logged-in person with a long chat history may see something different. We measure the questions you chose, so the question set shapes every figure. You add questions one by one, or accept the ones BotSay recommends from your website address.

Your first results arrive right after setup

Find the page to write next.

Monthly plans from $59. Cancel at any time.

Get started