ExplainerAI visibility
AI Search Share of Voice: How to Calculate It
by the AEO GEO Labs team6 min read
AI share of voice is the percentage of brand mentions in AI-generated answers that belong to your brand, out of all mentions of you and a defined set of competitors, across a fixed prompt set. To calculate it, run your prompts several times on each AI engine, count mentions per brand, then divide your count by the total. It tells you how much of the conversation you own.
Share of voice is one of the core numbers in AI visibility, and it's the one that puts every other metric in context.
The AI share of voice formula
The basic AI share of voice formula is your brand's mentions divided by total mentions of all tracked brands, times 100.
AI share of voice = (your mentions ÷ mentions of you + all tracked competitors) × 100
Three inputs decide what the number means:
- The prompt set. Which questions you test. A set full of branded prompts will inflate your share.
- The competitor set. Which brands count toward the total. Add or drop a competitor and every share changes.
- The runs. How many times each prompt runs on each engine. AI answers vary between runs, so one run per prompt produces a noisy share.
Fix all three before your first report and keep them stable. Changing any of them breaks comparison with earlier months.
A worked example of share of voice in AI search
Here is a worked example of share of voice in AI search, using made-up numbers for a project management tool called Acme.
The setup: 20 non-branded prompts, three engines (ChatGPT, Perplexity, Google AI Mode), three runs each. That's 180 runs. Four brands are tracked.
| Brand | Mentions across 180 runs | Mention rate | Share of voice |
|---|---|---|---|
| Acme | 63 | 35% | 63 ÷ 360 = 17.5% |
| Competitor A | 144 | 80% | 40.0% |
| Competitor B | 99 | 55% | 27.5% |
| Competitor C | 54 | 30% | 15.0% |
| Total | 360 | 100% |
Two things to notice. First, mention rate and share of voice answer different questions. Acme appears in 35% of answers, which sounds fine, but it holds only 17.5% of the conversation because answers usually name several brands. Second, shares always sum to 100% within the tracked set, so they show relative position, not absolute presence. Report both. The other metrics that sit alongside it are covered in AI visibility metrics.
Weighting options: count, position and citation
You can weight AI share of voice to reflect that some mentions are worth more than others. Each option adds realism and subjectivity in equal measure.
| Method | How it works | Use when |
|---|---|---|
| Unweighted (count) | Every mention counts as 1 | Default. Simple and hard to argue with |
| Position-weighted | First-named brand gets more weight, e.g. 1.0, 0.8, 0.6 | You care about being the top pick |
| Recommendation-weighted | Explicit recommendations count double | Buying-intent prompts dominate your set |
| Citation share | Your cited URLs ÷ all cited URLs | You care about sources and clicks, not names |
Our view: start unweighted. Position in AI answers is unstable, and SparkToro and Gumshoe found roughly a 1 in 1,000 chance of two answers listing the same brands in the same order. Weighting by position amplifies that noise. If you add weights later, publish the scheme in the report so readers can see what the number means.
Citation share is the variant with first-party data behind it. In June 2026, Microsoft added a Citation Share view to Bing Webmaster Tools, defined as the percentage of citations attributed to your site out of all citations shown for the same grounding query. It covers only Microsoft Copilot and Bing AI surfaces, but it's measured by the engine, not estimated.
How to calculate AI share of voice step by step
Calculate AI share of voice in six steps, by hand or with a tracking tool.
- Write 20 to 50 non-branded prompts your buyers ask: category, problem and comparison questions. Keep branded prompts in a separate set.
- Pick three to five competitors that buyers actually compare you with, not the biggest names in the industry.
- Choose engines your buyers use, and run each prompt at least three times per engine in clean sessions.
- Log every brand named in every run. Count each brand once per answer, however many times the answer repeats it.
- Calculate share per engine first, then overall. Engines differ, and a blended number can hide a big gap on one of them.
- Repeat monthly with the same prompts, competitors and engines. To connect share of voice to traffic and pipeline, see how to measure GEO results.
The logging template and session setup are in how to track AI brand mentions. If you track more than about 50 prompts, a tool will save time; the best AI visibility tools compares the options.
Common share of voice traps
The most common AI share of voice mistakes make the number look better or worse than reality.
- Branded prompts in the set. "Is Acme good for agencies?" will mention Acme every time. Keep those separate.
- Too few runs. Ten prompts with one run each is ten samples. A swing of one answer moves your share by several points.
- Missing competitors. If engines often name a brand you didn't track, your share is overstated. Review the "other brands" column monthly and add frequent ones.
- Blending engines. A 25% overall share could be 45% on Perplexity and 5% on ChatGPT. The fix for each is different.
- Reading AI visibility scores as share of voice. Some tools report a proprietary "AI visibility score". Check whether it's a share, a rate or a blend before comparing it with your own numbers.
Frequently asked questions
What is AI share of voice?
AI share of voice is the share of brand mentions in AI-generated answers that go to your brand, compared with a set of competitors, across a fixed list of prompts. If your brand gets 50 of 200 total mentions across ChatGPT, Perplexity and Google AI answers, your AI share of voice is 25%. It measures relative presence, not absolute reach.
How is AI share of voice different from an AI visibility score?
AI share of voice is a defined ratio: your mentions divided by all tracked brands' mentions. An AI visibility score is usually a proprietary number a tool calculates from several inputs, such as mention rate, position and citations, sometimes weighted by estimated prompt volume. Scores can be useful, but you should know how they're built before comparing them across tools.
What is a good AI share of voice?
A good AI share of voice depends on how many brands compete in your category. With four tracked brands, an even split is 25%, so anything above that means you're ahead of average. Compare your share with your market share: if you hold 30% of the market but 15% of AI mentions, engines are underrepresenting you, and that gap is the opportunity.
Should I count citations or mentions for share of voice?
Count mentions for your main share of voice figure, because brand names in the answer are what most readers see. Track citation share separately, because cited links drive clicks and show which of your pages engines trust. Bing Webmaster Tools reports citation share for Microsoft Copilot and Bing AI answers, which gives you one first-party benchmark.
If you want share of voice tracked against your competitors, AEO GEO Labs runs GEO, AEO and SEO programs for B2B and SaaS teams, starting with a free AI visibility report. See our services.