
Industries
AI visibility for professional services firms: get named
The questions clients ask AI assistants about law, accounting and consulting firms, what the answers say about you, and how to be the firm they name.
What share of voice counts in AI answers, how it differs from Mention rate, why it moves, and how to raise it.

Anyone who has tracked share of voice in search or social knows the idea: how much of the attention in a space goes to you. It carries over to AI answers. When a buyer asks an assistant a question, the answer usually names a few brands, and your share of voice tells you how much of that naming goes to you.
How we measure defines it in one line: of every brand mention in those answers, the share that was you. By "those answers" we mean the answers to the questions you track, on every engine you track.
In social listening, share of voice is your mentions out of all the mentions in a set of conversations. AI answers work the same way, since you are counting names, with one difference: you pick the questions, so the figure covers the topics you want to win.
Keep questions that name your brand in a separate count. An answer to "Is [your brand] any good?" names you almost by definition. Base your headline share of voice on the questions that leave your name out, because a mention there is one you earned.
Mention rate is the share of answers that named you, so it counts answers, while share of voice counts names. Take an answer that names you and four rivals. It counts once toward your Mention rate, and for share of voice it adds one mention of you and five to the total.
So the two can drift apart. If the answers that name you carry long lists, you can appear in many of them and still hold a modest share. A brand named in fewer answers, alone or beside one rival, can hold a large one.
The picture below shows every brand named in one week of answers for an invented brand, and the share each one holds. The invented brand holds the second-largest share, with one rival ahead of it and the rest split among several others.
A share like this means you are on the list but sharing it. The work then is to turn up in more of the answers where rivals appear.
Share of voice is a fraction, and either part of it can change from week to week.
This works in reverse too. A rival dropping out of answers lifts your share without a single new mention of you.
Share of voice rises when you start being named in answers that name your rivals today. Most of that comes from two kinds of work.
Rank each piece of work by how many of your questions it could win and start at the top. Note the day each piece goes live so you can match it to the readings that follow.
BotSay shows share of voice as this week's figure and how far it moved from last week. Counted from a few answers, a share can swing on a single reply, and it settles once the count reaches a few hundred.
So to judge a move in share of voice, read it beside Mention rate. Mention rate tells you whether you are named in more answers, and BotSay shows it with its likely range, the span the true figure almost certainly sits in, and the number of answers behind it. Here is an invented brand's Mention rate on two readings.
| Week | Mention rate | Likely range (%) |
|---|---|---|
| Last week | 27.8% | 25.1 to 30.7 |
| This week | 34.0% | 31.1 to 37.0 |
Example, using an invented brand
The two ranges sit clear of each other, so the rise is real. BotSay calls a change in Mention rate real only once it clears both weeks' uncertainty, and it runs the exact test, which also finds real changes whose ranges overlap a little. More answers give a narrower range, and our guide How many questions do you need to track AI visibility? shows how the number of questions you track sets it.
If Mention rate shows a real rise and share of voice rose with it, you won mentions. If share of voice rose and Mention rate held still, look at who left the answers before you report it. Either way, keep the questions, their wording and the engines the same from week to week, so each reading compares with the last. The full method is on How we measure.
Done by hand, the counting adds up fast. Ten questions on six engines is 60 answers a week, each with its own list of brands to tally. BotSay asks your questions on all six engines every week, through each provider's official developer access or a licensed search-data provider, and works out the five measures from the same answers. Your reading arrives in the weekly email.

Industries
The questions clients ask AI assistants about law, accounting and consulting firms, what the answers say about you, and how to be the firm they name.

Industries
The questions software buyers ask AI assistants, the measures that show whether you are on the shortlist, and a first month of work to be named more often.

Getting started
Run a hand check of AI answers this week, then track five numbers to see whether you are moving, and use the answers to decide what to write and fix.
Six engines on every plan
Monthly plans from $59. Cancel at any time.