Definition
AI share of voice (AISoV) is a metric that shows how much of the brand presence in AI-generated answers belongs to one brand, compared with its competitors. It’s calculated across a fixed set of prompts, the questions buyers put to tools such as ChatGPT, Gemini, Perplexity and Google’s AI Overviews.
The most common version counts brand mentions. If AI answers to a set of category questions name brands 150 times in total and 45 of those mentions are yours, your AI share of voice is 30%.
The name borrows from traditional share of voice, which measures a brand’s portion of advertising or media presence in a market. The AI version applies the same idea to a different surface: the text an AI system writes when someone asks it for a recommendation.
AI share of voice is closely related to Share of Model (SoM), and many people use the two interchangeably. Where a distinction is drawn, Share of Model is the broader concept, covering prominence and favorability as well as frequency. AI share of voice is the label most measurement tools put on the competitive percentage itself.
See also: Share of Model (SoM), Share of Voice (SOV), Generative Engine Optimization (GEO), Answer Engine Optimization (AEO)
Why it matters for marketing
Buyers increasingly ask an AI assistant which products to consider before they visit a vendor’s site. The answer usually names a handful of brands. A brand that isn’t named doesn’t get considered at that step, and it has no search ranking or ad impression to show for the miss.
Existing metrics don’t capture that. Search rankings report where a page appears in a list of links. Traditional share of voice reports ad spend or media mentions. Neither says whether ChatGPT includes the brand when a prospect asks for “the best marketing automation platforms for a mid-size B2B company.”
AI share of voice gives marketing leaders a competitive number for that surface. It can be tracked over time, compared across AI platforms and broken out by topic or buying stage. It’s also the figure most AI visibility tools now report by default, so it tends to be the one that ends up in a dashboard or board deck.
The metric matters most in categories where buyers research before purchasing: B2B software, financial services, healthcare, travel, and considered consumer purchases.
Hear it on The Agile Brand
These episodes of The Agile Brand with Greg Kihlström cover AI search visibility and how brands are measuring it.
- #801: Brandlight CEO Imri Marcus on GEO, AI browsers and the new search experience
- #844: adMarketplace’s John Nitti on GEO and the future of paid search
Find more episodes on The Agile Brand podcast page.
How to calculate AI share of voice
The basic formula counts mentions.
AI share of voice = (your brand’s mentions ÷ total mentions of all tracked brands) × 100
An illustrative example. A marketing automation vendor runs 100 buyer prompts through an AI assistant and counts every time a tracked brand is named.
| Brand | Mentions |
|---|---|
| Your brand | 45 |
| Competitor A | 60 |
| Competitor B | 30 |
| Competitor C | 15 |
| Total | 150 |
AI share of voice for the brand is 45 ÷ 150 × 100 = 30%.
The same data can give different numbers
There isn’t one agreed formula. Tools differ on what goes on top of the fraction and what goes underneath, so two tools can report different figures for the same brand on the same prompts.
| Variant | What it counts | Example result |
|---|---|---|
| Mention share | Your mentions as a share of all brand mentions | 45 ÷ 150 = 30% |
| Answer presence rate | Share of answers that name your brand at all | 45 of 100 answers = 45% |
| Citation share | Links to your domain as a share of all source links | 18 of 240 citations = 7.5% |
| Position-weighted share | Mentions weighted by how early the brand appears | Depends on the weights used |
| Impression-weighted share | Mentions weighted by estimated search volume of each prompt | Depends on volume estimates |
The example figures for citations are illustrative, like the rest of the table.
Published tool definitions show the spread. Ahrefs’ Brand Radar defines AI share of voice as a brand’s percentage share of impressions compared with other tracked brands, where impressions are estimated from the search volume of prompts in which the brand appears. Semrush’s reports factor in both how often a brand is mentioned and how high it appears in the answer. Other tools use a straight count of mentions.
Mentions and citations deserve separate attention. A mention means the brand is named in the answer text. A citation means the AI system linked to the brand’s website as a source. A brand can be mentioned often and rarely cited, or the reverse.
How to measure AI share of voice for a B2B brand
- Build the prompt set. Write the questions buyers ask at each stage: category questions (“what is a customer data platform”), shortlist questions (“best CDPs for retail”), and comparison questions (“Vendor A vs. Vendor B”). Sales call notes, search query data and customer interviews are good sources. Leave your brand name out of most prompts.
- Define the competitor set. List the brands you’ll count. Without competitors the metric is meaningless, since a brand measured alone always scores 100%.
- Choose the AI platforms. Cover the ones your buyers use. Results differ from one platform to the next.
- Run each prompt more than once. AI answers vary between runs, so a single response isn’t a reliable sample.
- Count mentions and citations separately. Record which brands are named and which domains are linked.
- Calculate, then segment. Work out the overall share, then break it down by platform, topic and buying stage.
- Repeat on a schedule. Use the same prompts each time so the trend is comparable.
For B2B brands, the segmented view is usually more useful than the total. A vendor might hold 35% of mentions on general category prompts and 5% on prompts about a specific industry it wants to grow in. That gap tells the content team where to work.
Small teams can do this by hand with 20 to 30 prompts in a spreadsheet. Larger programs use an AI visibility tool that runs the prompts automatically.
How to utilize AI share of voice
Competitive benchmarking. See which competitors AI systems recommend most often in the category.
Finding content gaps. Identify the prompts where competitors appear and the brand doesn’t, then create or improve content on those topics.
Tracking GEO work. Measure whether generative engine optimization efforts change how often the brand is named.
Platform comparison. Find out whether the brand is strong in one AI system and missing from another.
PR and source strategy. Look at which third-party sites AI systems cite in the category, since coverage on those sites can influence answers.
Executive reporting. Give leadership one competitive figure for AI visibility alongside search and media metrics.
Launch monitoring. Check whether a new product or repositioning shows up in AI answers after launch.
Comparison with related metrics
| Metric | What it measures | Where | Basis |
|---|---|---|---|
| AI share of voice | A brand’s share of mentions or citations in AI answers | AI assistants and AI search | Sampled prompts |
| Share of Model (SoM) | How often, how prominently and how favorably a brand appears in AI answers | Large language models | Sampled prompts |
| Share of Voice (SOV) | A brand’s share of advertising or media presence | Paid, owned and earned media | Spend, impressions or mentions |
| Share of search | A brand’s share of branded search queries in its category | Search engines | Search volume |
| Share of Market (SOM) | A brand’s share of category sales | The market | Revenue or units |
| AI visibility score | A blended score combining mentions, citations, sentiment and other signals | AI assistants and AI search | Vendor-specific formula |
| Citation share | How often a brand’s domain is linked as a source | AI answers with sources | Sampled prompts |
One practical difference separates AI share of voice from the traditional kind. Traditional share of voice can be bought, because more ad spend produces more impressions. AI share of voice can’t be purchased directly in most AI answers. It’s earned through the content and third-party sources that AI systems draw on.
The abbreviation SOM causes confusion, since it’s used for both Share of Model and Share of Market. Spell the term out in reports.
Best practices
Write the definition down. Record which formula, prompts, competitors and platforms are used, and keep them fixed.
Don’t compare numbers across tools. Different formulas produce different results from the same answers.
Report mentions and citations separately. They measure different things and often move independently.
Use unbranded prompts. Prompts that include the brand name inflate the score.
Sample repeatedly. Run each prompt several times to account for variation in AI answers.
Segment the results. A single overall figure hides the differences between platforms, topics and buying stages.
Check what’s being said. A high share is less useful if the brand is described inaccurately or unfavorably.
Connect it to outcomes. Track AI referral traffic and pipeline alongside the share figure.
Expect movement. AI model updates can shift results without any change in the brand’s own activity.
Limitations
No standard definition. The metric is new, and vendors calculate it in different ways.
Prompt-set dependence. The result reflects the prompts chosen. A different set gives a different number.
Variable answers. AI systems can respond differently to the same prompt from one run to the next.
Personalization. Real users’ answers may be shaped by their history or location, which a test prompt doesn’t reproduce.
Unknown real-world volume. AI platforms don’t publish how often each question is asked, so weighting by demand relies on estimates.
Unproven link to revenue. The relationship between AI share of voice and sales hasn’t been established the way the link between traditional share of voice and market share has been studied.
Future trends
More data from the platforms. Google’s Merchant Center reportedly offers a native share of voice figure based on AI impressions. First-party reporting of this kind would reduce reliance on sampled prompts.
Larger measurement panels. Semrush’s AI Visibility Index reportedly grew from 2,500 prompts to 126 million between versions. A jump of that size shows how quickly measurement is scaling.
Paid placements in AI answers. As advertising enters AI assistants, tools will need to separate earned mentions from paid ones.
Agentic commerce. When AI agents make purchases on a buyer’s behalf, being selected matters more than being mentioned, and metrics are likely to follow.
Pressure to standardize. As the metric appears in more executive reports, demand for a consistent definition is growing.
Integration with existing dashboards. SEO suites have added AI share of voice next to traditional rankings, so it’s increasingly tracked in the same place.
Frequently asked questions
What is AI share of voice? It’s the percentage of brand mentions in AI-generated answers that belong to your brand, measured against competitors across a set of prompts.
How do you calculate AI share of voice? Divide your brand’s mentions by the total mentions of all tracked brands across the same AI answers, then multiply by 100.
How do you measure AI share of voice for a B2B brand? Build a set of buyer questions for each stage of the purchase process, pick the competitors and AI platforms to track, run the prompts repeatedly, and count how often each brand is named. Segment results by buying stage and industry.
Is AI share of voice the same as Share of Model? They overlap heavily and are often used as synonyms. Share of Model is usually defined more broadly, to include prominence and favorability. AI share of voice usually refers to the competitive percentage of mentions.
How is it different from traditional share of voice? Traditional share of voice measures a brand’s portion of advertising or media presence. AI share of voice measures its portion of mentions in answers written by AI systems.
What’s a good AI share of voice? There’s no reliable universal benchmark. It depends on the number of competitors and the prompts used. The trend over time and the gap to named competitors are more useful than the absolute figure.
What’s the difference between a mention and a citation? A mention is the brand being named in the answer. A citation is the AI system linking to the brand’s website as a source.
Which tools measure AI share of voice? SEO suites such as Semrush and Ahrefs report it, as do specialized AI visibility tools such as Profound. HubSpot offers a free grader. Each uses its own method.
Why do two tools show different numbers for the same brand? They use different prompts, different formulas and different AI platforms. Some count mentions, some count citations, and some weight by position or search volume.
How often should it be measured? Monthly is common for manual tracking. Automated tools often run weekly or daily. Consistency of prompts matters more than frequency.
Related terms
- Share of Model (SoM)
- Share of Voice (SOV)
- Share of Market (SOM)
- Generative Engine Optimization (GEO)
- Answer Engine Optimization (AEO)
- Citation Optimization
- AI Overviews
- llms.txt
- Large Language Models (LLM)
- Search Engine Optimization (SEO)
Sources
- Ahrefs Help Center. “AI Visibility metrics.” https://help.ahrefs.com/en/articles/15501968-ai-visibility-metrics
- Hartzer Consulting. “Share of Voice in AI Search.” https://hartzer.it.com/share-of-voice-in-ai-search/
- Exploding Topics. “Share of Voice.” https://explodingtopics.com/blog/share-of-voice
- GEO Toolbox. “AI Share of Voice: How to Measure Your Slice of AI Answers.” https://geotoolbox.ai/blog/ai-share-of-voice
- LLM Pulse. “How to Measure AI Share of Voice (Complete Guide for 2026).” https://llmpulse.ai/blog/measure-ai-share-of-voice/
- AI Growth Agent. “AI Share Of Voice: Tools, Metrics & How To Act On It.” https://aigrowthagent.co/articles/?p=6311
- Massive. “What Is AI Share of Voice?” https://www.joinmassive.com/glossary/what-is-ai-share-of-voice
- The Agile Brand Guide. “Share of Model (SoM).” https://www.agilebrandguide.com/wiki/agentic-commerce/share-of-model-som/