AI share of voice is the percentage of competitive brand mentions in AI-generated answers that name your brand, measured across a fixed set of prompts. Share of model is the same calculation for a single AI model, and citation share is the percentage of cited sources that point to brand-owned pages. Each metric uses a different denominator, so each answers a different question.
Last updated: October 28, 2026
Understanding brand presence across modern digital channels requires distinguishing between three forms of AI visibility and traditional media metrics:
- AI Share of Voice: Measures a brand's percentage of total competitive brand mentions within AI-generated responses across a defined prompt set.
- Citation Share: Measures the percentage of cited sources pointing directly to brand-owned web properties.
- AI Visibility Rate: Measures the percentage of total AI responses where a brand is mentioned at least once, evaluating reach regardless of mention frequency.
This guide also defines share of model, which is AI share of voice calculated for one model at a time, and it shows how 5W's AI Visibility Index figures map to each formula.
Each metric measures a different position in the model's information ecosystem.
A brand may lead one metric while often trailing another, so brands that combine all three risk drawing the wrong conclusion from the data.
What Is Traditional Share of Voice?
Traditional share of voice measures a brand's share of media mentions relative to a defined set of competitors.
For example, consider Nike, Adidas, Puma, New Balance, and Under Armour. If these brands collectively generate 10,000 media mentions in a quarter and Nike accounts for 2,500, Nike holds a 25% share of voice in that category.
How Much of the Total Category Conversation Does a Brand Control?
That is the question traditional share of voice answers. PR teams use share of voice to benchmark their media presence against competitors and to track visibility changes after product launches, campaigns, or major news events.
However, traditional share of voice doesn't reveal how often AI models recommend a brand or whether they cite the brand's website as a source or recommendation in their answers. In short, traditional share of voice reflects media coverage, not AI model discovery.
What Is AI Share of Voice?
AI share of voice relies on a specific dataset built from targeted AI model responses.
Brand teams define relevant consumer queries and target large language models, markets, and evaluation parameters. They then analyze every brand mention that appears across the models generated responses.
AI share of voice represents a brand's percentage of total competitive brand mentions in AI-generated answers.
AI visibility platforms monitor this metric to evaluate brand prominence. Specifically, AI share of voice measures "how often a brand appears in AI responses across a prompt set relative to competitors, defining a response as the answer a model returns for a given prompt at a given point in time."
How Is AI Share of Voice Calculated?
For example, imagine analyzing AI responses to queries about "best running shoes," which generate 1,000 total brand mentions across all results. If Nike is mentioned 280 times, Nike holds a 28% AI share of voice for that query set.
How Much of the Brand Conversation Generated by AI Belongs to a Brand?
That is what AI share of voice answers. It helps brands compare visibility across competitors, consumer needs, product categories, and stages of selection.
Brands like Optimum Nutrition hold a strong share among queries about "whey isolate for muscle building" while getting little attention for "plant-based protein for sensitive stomachs." A single category-wide percentage could hide the difference between those two consumer needs.
Consequently, testing design matters as much as the final percentage.
AI share of voice isn't a universal market-share number. The percentage belongs to a defined test. Google states that AI Overviews and AI Mode may use different models and techniques, so the responses and supporting links vary between the two experiences. Google Search Central: AI Features and Your Website
AI share of voice fluctuates when the prompt set, model mix, geography, product category, timeframe, or competitive set changes. Brands should establish fixed parameters before comparing results over time.
What Is Share of Model and How Does It Differ From AI Share of Voice?
Share of model is a brand's share of competitive brand mentions within one named AI model, such as ChatGPT or Gemini, across a fixed prompt set. Vendors use the term differently, so this guide defines it by its denominator: all brand mentions in that one model's answers.
AI share of voice pools every tested model into one number, and share of model splits the same data by model. Splitting matters because models answer differently. 5W's banking research found meaningful platform differences, so a single-model view can mislead, and a pooled number can hide a weak model.
| Metric | Formula |
|---|---|
| AI share of voice | Brand mentions across all tested models ÷ all brand mentions across all tested models × 100 |
| Share of model (for model M) | Brand mentions in model M's answers ÷ all brand mentions in model M's answers × 100 |
The figures below are an illustration, not company data. A brand is named 120 times in 1,000 total brand mentions in one model's answers, which is a 12% share of model. In a second model it is named 60 times in 800 mentions, which is 7.5%. Pooled across both, it is named 180 times in 1,800 mentions, so its AI share of voice is 10%, and the pooled number hides the gap between the two models.
How Do 5W AI Visibility Index Numbers Map to These Formulas?
5W's AI Visibility Index reports "citation share" as a brand's percentage of all brand mentions inside AI-generated answers, so under the definitions in this guide those figures are AI share of voice, because the denominator is all mentions rather than cited sources. The Index publishes them as modeled, directional estimates and refreshes them quarterly.
The Theme Parks AI Visibility Index (May 2026) covers the top 25 destination parks across five AI answer systems and 50 traveler-intent prompts, and it shows how a share changes with the prompt set and the level of roll-up.
| Brand or group | Share reported | Prompt set |
|---|---|---|
| Disney properties combined | 58.1% | All 50 prompts |
| Comcast NBCUniversal properties combined | 23.5% | All 50 prompts |
| Six Flags Entertainment Corporation properties combined | 12.8% | All 50 prompts |
| Top three parents combined | 94.4% | All 50 prompts |
| Walt Disney World | 22.6% | All 50 prompts |
| Walt Disney World | 32.6% | 10 family-destination prompts |
| Cedar Point | 4.1% | All 50 prompts |
| Cedar Point | 26.8% | 10 thrill prompts |
Source: 5W Theme Parks AI Visibility Index, May 2026, modeled directional estimates.
The table shows two things. First, the same park holds a very different share on a different prompt set: Walt Disney World moves from 22.6% to 32.6%, and Cedar Point moves from 4.1% to 26.8%, which is why a share is only meaningful with its prompt set attached. Second, roll-up changes the number: the three parent companies sum to 94.4%, while the largest single park holds 22.6%.
The Banks AI Visibility Index gives a second example. It reported estimated AI citation shares of 9.2% for Ally, 7.8% for JPMorgan Chase, 6.5% for SoFi and 5.4% for Marcus by Goldman Sachs across 66 consumer-intent prompts in Q2 2026. These are brand-mention shares, so they are AI share of voice figures.
What Is Citation Share in AI Search?
Citation share tracks the percentage of citations pointing to brand-owned web properties within a given query set. These include assets under direct corporate control, such as product pages, technical documentation, newsroom announcements, and official FAQs.
Crucially, never count third-party sources toward brand-owned citation share, even if they feature a brand.
Relying on external links creates a misleading perception of source ownership. For example, if an AI model links to a retail page on Bloomingdale's to support a product recommendation, Bloomingdale's controls the page content, host domain, and contextual messaging, not the featured brand.
Similarly, when an industry publication cites a company's research or product line, the publication retains ownership of the source. While these third-party links support overall brand visibility, they don't count as brand-owned citation share because the brand can't control or update the content on those pages.
Distinguishing brand-owned citations from earned third-party citations provides a clearer picture of an organization's true authority and direct footprint within AI answer engines.
What Is the Citation Share Formula?
Brand-owned citation share equals the citations that point to brand-owned web properties, divided by all citations in the answers for the prompt set, times 100. The related source concentration measure uses the same denominator for third-party publishers: citations from the top three publishers, divided by all citations, times 100.
The figures in this paragraph are an illustration, not company data. If AI answers for a prompt set contain 400 citations and 52 point to the brand's own domains, brand-owned citation share is 13%.
5W's Banking AI Visibility Index shows the same formulas with real data. Across 31,500 prompts and five AI assistants from January to May 2026, Bankrate, Investopedia and Wikipedia supplied 68% of banking-related AI citations, and bank-owned domains accounted for less than 7%, according to 5W's banking AI search analysis. The 68% is a source concentration figure, and the under 7% is the combined brand-owned citation share for banks.
| Metric | Formula | 5W Index example | Question it answers |
|---|---|---|---|
| AI share of voice | Brand mentions ÷ all brand mentions × 100 | Ally 9.2% estimated share in the banking index; Disney properties 58.1% in the theme parks index | How much of the AI answer names the brand? |
| Share of model | Same formula, one model at a time | Not published by model in the reports cited here | Does each model name the brand equally? |
| Brand-owned citation share | Citations to brand-owned pages ÷ all citations × 100 | Bank-owned domains under 7% of banking citations | Do AI answers link to the brand's own pages? |
| Source concentration | Citations from top three publishers ÷ all citations × 100 | Bankrate, Investopedia and Wikipedia at 68% | Which third parties shape the answer? |
The banking figures come from two separate 5W studies with different prompt sets and dates, so they are not directly comparable. Taken together they show that the metrics measure different things: a bank can hold the largest share of mentions in an answer while all bank-owned domains together hold a small share of the citations behind it.
How Much of the Cited AI-Information Ecosystem Does a Brand Control?
That is what citation share reveals. While traditional share of voice evaluates earned media conversation, AI metrics measure different aspects of large language model (LLM) discovery: AI share of voice tracks dominance within AI-generated answers, citation share measures direct source authority, and visibility rate evaluates broad reach.
Search rankings add another measurement that should remain separate from citation share.
Ahrefs analyzed 863,000 keyword search results and four million URLs cited in Google AI Overviews in a March 2026 study. The analysis found that 37.9% of URLs cited in AI Overviews also appeared within the first 10 blocks of the search results for the same query.
Search position and AI citation presence therefore measure different forms of visibility.
Brands need these metrics because each question points toward a different action.
For low traditional share of voice, increase public relations outreach, launch targeted campaigns, or secure executive interviews in trade publications.
For low AI share of voice, expand content to cover long-tail consumer queries, create comparative 'versus' pages, and specific FAQs that address specific buying scenarios.
For low brand-owned citation share, publish technical documentation, detailed product description pages (PDPs), and official brand research that models scrape as expert primary sources for AI models.
One percentage can't diagnose all three problems.
Citation share also requires measurement over time. Another study analyzed more than 43,000 queries, each with at least 16 recorded AI Overviews during a month. The study found that 45.5% of cited URLs changed between consecutive observations.
A citation-share measurement taken once describes a snapshot without establishing a stable pattern.
Which AI Visibility Metric Should Brands Use?
Successful measurements start with goals, not with standard reporting dashboards. Every team needs specific metrics to answer the questions that drive revenue to their work:
- PR & Communications (Traditional Share of Voice): Focuses on earned media prominence. For instance, after launching a big product, a PR team looks at traditional share of voice to see if they won the news cycle over their competitors.
- Brand Marketing (AI Share of Voice): Measures how often you show up inside AI answers. Brand managers track this to ensure their product gets recommended right alongside top competitors when people ask AI tools for suggestions.
- Technical SEO & Content (Citation Share): Tracks direct source authority. Instead of hunting for press mentions, SEO specialists watch citation share to make sure AI models link straight to your own spec sheets, docs, and FAQs instead of third-party review sites.
- Growth & Awareness Teams (AI Visibility Rate): Evaluates overall reach and discovery. Growth leaders check this to see how consistently their brand pops up across key prompt sets, ensuring it shows up at least once in responses.
Combining each number into a catch-all score hides the actionable insights you need for a successful outcome.
Try this smart approach instead: pinpoint your strategic questions first, build the right dataset second, and pick the right metric third.
How to Use AI Share of Voice With Other AI Visibility Metrics
Right now, no single reliable metric (like a domain authority score) tells a brand about its overall AI visibility.
Traditional share of voice tracks media news coverage. AI share of voice measures how often you appear in AI answers compared to your competitors. Citation share shows how many source links point back to your website.
Each metric highlights a specific growth opportunity.
When brands use clear, distinct definitions, your marketing efforts work together effectively. Press coverage builds overall authority, helpful website content provides AI models accurate information, and targeted search analysis shows where brands stand against competitors.
Which Other AI Visibility Metrics Should Brands Track?
Brands should also track Recommendation Share, Source Citation Rate, Accuracy Rate, Query Coverage Rate and Cross-Platform Consistency Rate, because AI share of voice, citation share and AI Visibility Rate do not show whether a platform recommends the brand, supports it with a source, describes it correctly, or repeats the result across platforms. The AI Visibility Rate described above matches what some frameworks call Brand Mention Rate.
| Metric | How it is calculated | Question it answers |
|---|---|---|
| Recommendation Share | The brand's qualifying recommendations divided by all qualifying recommendations in the defined competitive group, times 100 | How often does AI recommend the brand instead of competitors? |
| Source Citation Rate | Qualifying brand appearances with a traceable supporting source divided by all qualifying appearances reviewed, times 100 | What evidence supports the brand's visibility? |
| Accuracy Rate | Brand appearances with no material factual error divided by all appearances reviewed, times 100 | Is AI describing the brand correctly? |
| Query Coverage Rate | Priority prompts that produce at least one brand appearance divided by all priority prompts tested, times 100 | Which customer questions does the brand currently answer? |
| Cross-Platform Consistency Rate | The average, across priority prompts, of the share of tested platforms that name or recommend the brand | Does visibility repeat across platforms or depend on one? |
Source Citation Rate and citation share are different measures. Source Citation Rate counts any traceable supporting source behind a brand appearance, including third-party pages, while citation share counts only sources the brand owns.
Recommendation Share measures AI-layer visibility only. It does not measure revenue, customers or financial market share, and each qualifying recommendation should be defined in advance, for example as a brand surfaced as a top recommendation within a specific category and geographic market.
How Does Accuracy Rate Change the Reading of the Other Metrics?
Accuracy Rate changes the reading of every other metric, because a brand mentioned often for the wrong reasons has a reputation problem, not a visibility gain. A review should count an answer as inaccurate if it contains at least one material factual error, such as the wrong industry category, an outdated price, a former executive or an invented partnership.
In an illustrative example, a brand appears in 40 AI answers and 8 of them contain at least one material error. That leaves 32 accurate appearances, so Accuracy Rate is (32 ÷ 40) × 100 = 80%. The figures are an example, not a company finding.
How Should Brands Set an AI Visibility Baseline?
A brand should record a baseline before starting any GEO, SEO or PR work, so later changes can be separated from platform and model updates. The baseline documents the exact prompt set, the platforms tested, the model or mode shown, the test date and location, the competitors included, and the rules used to score recommendations and accuracy.
Re-run the same prompts quarterly or at another fixed interval. The Ahrefs study of more than 43,000 queries cited above found that 45.5% of cited URLs changed between consecutive observations, so a single test describes one moment, not a trend.
A report should show each metric separately and by platform, along with the prompt wording and test conditions. Percentages used as examples should be labeled as illustrations unless a documented audit produced them. 5W's AI Visibility Index documents its prompts, platforms and testing dates in the methodology that accompanies each report.
Frequently Asked Questions
What is share of model?
Share of model is a brand's share of competitive brand mentions within one named AI model across a fixed prompt set. It uses the same formula as AI share of voice, calculated for one model at a time instead of pooled across all tested models.
What is the formula for citation share?
Brand-owned citation share is the citations pointing to brand-owned web properties, divided by all citations in the answers for the prompt set, times 100. Third-party pages that mention the brand do not count toward the numerator.
How is AI share of voice different from citation share?
AI share of voice counts brand mentions in AI answers against competitors, and citation share counts how many cited sources are the brand's own pages. A brand can lead one and trail the other, as the 5W banking studies show.
Is AI share of voice the same as market share?
AI share of voice is not market share. It describes how often a brand appears in a defined set of AI answers, and it says nothing about revenue, customers or product quality.
How often should brands re-measure AI visibility?
Brands should re-run the same prompts on a fixed schedule, such as quarterly. An Ahrefs study found that 45.5% of cited URLs changed between consecutive observations, so a single measurement describes one moment, not a trend.
How 5WPR Connects AI Share of Voice With PR and AI Search
At 5WPR, we combine strategic public relations with advanced AI search optimization to turn complex data into clear business outcomes. We help you identify your visibility baseline, outperform competitors in AI recommendations, and drive qualified traffic to your brand-owned channels.
