On this page
The short version
AI share of voice is your share of all the businesses AI names when people ask your customers' questions. It's useful, but it's only half the picture, and it can point the opposite way to the number your client actually feels.
- The formula: your mentions ÷ all business mentions. For Kato Electrical, that's 302 ÷ 2,602 = 11.6%. (Formula)
- The other half: the named rate, answers naming you ÷ all answers. For Kato, 17.0%. (Named rate)
- They can disagree. Perplexity named Kato in 23.9% of its answers but gave him only 9.6% share of voice, because it lists so many businesses. (By AI system)
- Kato leads his market's share of voice at 11.6%, nearly double the next electrician, and is still named in fewer than 1 answer in 5. (Leaderboard)
- How to report it honestly: per AI system, with sample sizes, merged spellings and both numbers. (Reporting)
What AI share of voice is
The short answer: of all the businesses AI mentioned for your questions, the share that were you.
Share of voice is an old marketing idea: your share of the conversation in your category, whether that's ad spend, press mentions or social posts. In AI search, the "conversation" is the set of businesses AI names when someone asks a buying question.
The formula:
AI share of voice = mentions of your business ÷ mentions of all businesses, across the same questions, AI systems and days.
The worked example is Kato Electrical, a Vancouver-area electrician and our public demo. Our tracker, LLM Visibility Lens, asked his buying questions every day in ChatGPT, Gemini, Perplexity, Google AI Mode and Google AI Overviews, from 30 September to 21 October 2026. Across 1,781 answers, AI mentioned a business 2,602 times. Kato was 302 of those mentions, so his share of voice is 11.6%.
I sell this tracker, so weigh my view. The formula works with a spreadsheet too.
Share of voice vs named rate
The short answer: share of voice measures your slice of the conversation; named rate measures how often a customer actually sees your name.
| Share of voice | Named rate | |
|---|---|---|
| Question it answers | How much of the conversation is about you? | How often does a customer see your name? |
| Formula | Your mentions ÷ all business mentions | Answers naming you ÷ all answers |
| Kato | 302 ÷ 2,602 = 11.6% | 302 ÷ 1,781 = 17.0% |
| Goes up when | Competitors are named less | You're named more often |
Both matter. A client cares most whether a customer sees their name, which is the named rate. Share of voice adds the competitive picture: how much room competitors are taking up.
Some tools also call the named rate "visibility" or "mention rate". Whatever the label, ask which of the two you're looking at. We've seen the confusion in client reports; see why a single AI visibility score can mislead you.
Where the two numbers disagree
The short answer: AI systems that name long lists dilute everyone's share of voice, even when they name you often.
| AI system | Kato's share of voice | Kato's named rate |
|---|---|---|
| ChatGPT | 7.8% | 14.7% |
| Gemini | 13.0% | 10.6% |
| Google AI Mode | 18.4% | 23.4% |
| AI Overviews | 15.0% | 11.3% |
| Perplexity | 9.6% | 23.9% |
The sharpest case is Perplexity:
- It named Kato in 23.9% of its answers, the highest of all five.
- But it made 912 business mentions in total, so Kato was only 9.6% of the conversation.
Gemini and AI Overviews go the other way. They name fewer businesses per answer, so when Kato appears he takes a bigger slice: share of voice is higher than named rate.
So a report showing only share of voice would call Perplexity one of Kato's weakest AI systems. A report showing only named rate would call it his strongest. Both are true, which is why you show both. We compared how many businesses each AI names in which AI system recommends local businesses most.
The competitor leaderboard
The short answer: share of voice is most useful as a leaderboard, showing who takes up the conversation and by how much.
| Business named by AI | Mentions | Share of voice |
|---|---|---|
| Kato Electrical | 302 | 11.6% |
| Intel Power Electric | 165 | 6.3% |
| Fusion Point Electric | 127 | 4.9% |
| TDR Electric | 111 | 4.3% |
| Expert Electric | 74 | 2.8% |
| PowerX Electrical | 65 | 2.5% |
| Electric Hero | 62 | 2.4% |
| Whitley Electric | 62 | 2.4% |
| Arcwest Electric | 62 | 2.4% |
These are businesses AI named in its answers, not a ranking of who's best. Kato leads, with nearly double the next electrician's share. Dozens of other businesses share the remaining 60%, which says something on its own: AI spreads recommendations widely in a city this size.
And yet Kato is named in fewer than one answer in five. Leading the leaderboard doesn't mean customers see you most of the time; it means you're the biggest name in a crowded, fragmented conversation. Where he's missing is the useful part: Coquitlam and the long spoken questions, as we showed in the 30-day plan for when AI ignores a client.
Traps to avoid
The short answer: most share of voice numbers go wrong in five predictable ways.
- Name spellings. "Intel Power Electric Ltd.", "Intel Power Electric" and "Intel Power Electrical" are probably one company. We merge obvious variants like "Ltd." and "Inc.", but similar names can still split. Check the top of your leaderboard by eye.
- Brand questions. Questions with your client's name in them inflate share of voice. In one test, a brand was named in 88% of answers to questions containing its name and 0% of ready-to-buy questions. Use questions a new customer would ask.
- Changing questions. Add or drop questions and share of voice moves for no real reason. Compare periods on the same set only.
- Small samples. AI answers vary day to day; five AI systems agreed only about 1 time in 5 on the same morning (5 AIs, same questions). SparkToro's research found the same inconsistency. Report the number of answers behind every figure.
- Blending AI systems without showing them. One overall figure hides that Kato's share of voice ranges from 7.8% in ChatGPT to 18.4% in Google AI Mode.
One more: when Google shows no AI Overview for a search, there's no answer to count. Leave those searches out rather than counting them as zero. Google explains when AI features appear in AI features and your website.
How to report AI share of voice
The short answer: one table per period, with both numbers, per AI system, and the sample behind each.
| Section | What to show | Kato's example |
|---|---|---|
| Headline | Share of voice and named rate, with answer counts | 11.6% of 2,602 mentions; named in 17.0% of 1,781 answers |
| By AI system | Both numbers for each | AI Mode 18.4% / 23.4%; Perplexity 9.6% / 23.9% |
| Leaderboard | Top 5–10 businesses AI named, merged spellings | Kato 11.6%, Intel Power 6.3%, Fusion Point 4.9% |
| Change | Same questions vs last period, in points | Only when the question set is unchanged |
| Where to win | Questions and towns with the lowest share | Coquitlam and long spoken questions |
| Method | Questions, AI systems, dates, how spellings were merged | 30 Sep–21 Oct 2026, five AI systems |
The full client report format is in how to report AI visibility to clients.
What I tell clients: "Of all the electricians AI mentioned for your customers' questions, 11.6% of the mentions were you, more than anyone else. But a customer asking those questions sees your name in about 1 answer in 6. Our job is to raise that second number."
How we track it
LLM Visibility Lens asks your questions in all five AI systems daily, counts every business named, and shows your named rate and the competitors AI named, per AI system, with answer counts. Agencies get the report under their own name; see the agency plan. See it on Kato's account in the live demo, or compare plans, which start free.
More reading:
- Which websites ChatGPT trusts
- The directories AI reads
- Which pages AI reads
- Questions nobody searches on Google
- The AI visibility checklist
- Invisible on Google, recommended by AI
- When your own web page helps a competitor
- How to get recommended by ChatGPT
- 5 AI systems, 35 questions
- Google Business Profile and AI answers
- Google AI Mode for local businesses
- How to rank in AI Overviews
- AI Overview tracking for near me searches
- Perplexity vs ChatGPT
- ChatGPT citations
- Spoken questions
- How to rank in ChatGPT
- Why there's no rank #1 in ChatGPT
- What one free check can and can't do
- We asked AI who to hire 80 times
The method is on the how it works page. There's also a comparison with Semrush's AI visibility tool, the changelog, the FAQ page, the security page and our free question finder. If you'd like help setting up share of voice reporting for a client, get in touch.
Questions people ask
Harjeet Singh Bedi Founder, LLM Visibility Lens · SEO consultant, Bedi SEO
Harjeet has worked in SEO since 2010, first in agencies and now running Bedi SEO for local businesses, online stores and agencies. He built LLM Visibility Lens after clients kept asking whether they showed up in ChatGPT and other AI tools and nobody could give them an honest, measured answer. Every figure in this article comes from answers his tool collected.