What's AI share of voice and how is it measured?

AI share of voice is the percentage of AI assistant answers about your category that name your brand. It is measured by sampling: fix a set of real buyer questions, run each one repeatedly across ChatGPT, Claude, Gemini, and Perplexity, and count the share of answers that name you rather than a competitor.

Table of Contents

• How do I measure my brand's AI share of voice?

• The four-step method, in short

• What the numbers look like across our own measurement

• Why does Share of Voice matter for B2B marketing?

• How do AI providers differ in how they cite brands?

• What is the difference between AI visibility and SEO?

• How do I improve my AI mention rate in 30 days?

• FAQ

How do I measure my brand's AI share of voice?

To measure AI share of voice, you run Discovery, Evaluation, and Decision prompts across ChatGPT, Claude, Gemini, and Perplexity, then calculate the percentage of answers that name your company against the answers that name competitors instead. Alongside the raw percentage you track where in the answer you appear, the sentiment of the description, and which sources the model cited to support it.

Measuring this manually is impossible because model output varies between runs. A brand might appear in ChatGPT for a "best project management software" query in 40% of sessions but only 10% in Perplexity, and a single run of one prompt tells you nothing you can act on. Any number worth reporting comes from repeated runs across every assistant your buyers use.

A significant and growing number of B2B buyers now use AI assistants to conduct initial vendor research before visiting a company website.

When measuring, you must categorize prompts by intent:

1. Discovery: "What are the top tools for SOC2 compliance?"
2. Evaluation: "Compare Vanta vs. Drata for a Series B startup."
3. Decision: "What are the main complaints about Vanta's automated evidence collection?"

The four-step method, in short

Monroya publishes the measurement framework it uses, called AI Share of Voice (ASOV). The short version is four steps:

1. Frame the question set. Write the questions your buyers actually ask, split across Discovery, Evaluation, and Decision, rather than the keywords you wish you ranked for.
2. Sample across models and runs. Ask every question many times on each assistant, on a fixed cadence, so the result reflects the model's behaviour rather than one lucky answer.
3. Score presence, not raw mentions. A recommendation is worth more than a passing name-check, so answers are scored by how you appear, not only whether you appear.
4. Attribute to sources and act. Record which pages each answer cited, so the fix is a specific page rather than a vague instruction to publish more.

The full framework, including the formula and a worked example, is on the AI Share of Voice page. How each score is calculated is written up in our measurement methodology.

What the numbers look like across our own measurement

These figures come from Monroya's own scans, not from a third-party study.

• Between June 16 and September 17, 2026, Monroya recorded 182,517 assistant answers across 48,559 tracked questions and four assistants (Monroya, September 2026).
• 54.3% of those answers named at least one brand, which means close to half of all answers in these categories named no vendor at all (Monroya, September 2026).
• Across the 158 companies in Monroya's 2026 AI visibility benchmark, measured in September 2026, the average AI readiness score is 50.9 out of 100, and 17 companies, 11% of the cohort, score below 40 and fall into the Not Ready tier.

The gap between those numbers is the opportunity: in roughly half of category answers, no vendor is named, and most measured companies are not structured well enough to be the one that is.

Why does Share of Voice matter for B2B marketing?

Share of voice is the primary indicator of whether your brand exists in the answer space your buyers use to build shortlists. If a buyer asks Claude for a recommendation and your brand is never mentioned, you have no chance of winning that deal, because you were never in the consideration set. High share of voice correlates with increased organic branded search and direct traffic.

In 2026, the traditional marketing funnel has been compressed. Buyers no longer click through ten blue links; they ask Perplexity to summarize the market. If your share of voice is low, your top-of-funnel demand generation is effectively broken.

Companies with a high Share of Voice relative to their market share often experience a corresponding increase in pipeline velocity.

How do AI providers differ in how they cite brands?

AI providers use different retrieval mechanisms: Perplexity and Gemini rely heavily on real-time web indexing and recent news, while ChatGPT and Claude lean more on their training data and specific high-authority datasets like Reddit or GitHub. Understanding these differences allows you to tailor your Generative Engine Optimization (GEO) strategy to the specific model where your buyers spend the most time.

For example, if you are a developer-tooling company, your visibility in Claude is often tied to your documentation's presence in the training set. Conversely, your visibility in Perplexity is tied to recent mentions in technical forums and press releases.

Perplexity citations are significantly more likely to come from community-driven sites like Reddit and Stack Overflow than from corporate blogs.

What is the difference between AI visibility and SEO?

SEO focuses on optimizing for a search engine's ranking algorithm to drive clicks, whereas AI visibility focuses on optimizing for a model's probability of selecting your brand as the most relevant answer. SEO is about being found; AI visibility is about being recommended. While SEO relies on keywords, AI visibility relies on "entities" and the relationships between them in the model's neural network.

If you want to understand how your brand compares to others in this new landscape, you can look at tools like Profound vs monroya.ai. While some tools focus on general search, monroya.ai specifically tracks the B2B Buyer Journey across ChatGPT, Claude, Gemini, and Perplexity.

How do I improve my AI mention rate in 30 days?

To improve your AI mention rate, you must first identify the "knowledge gaps" where models are hallucinating or ignoring your brand, then fill those gaps with high-authority, structured data. This includes updating your FAQ schema, increasing your presence in third-party review sites, and ensuring your technical documentation is easily crawlable by LLM bots.

1. Audit your current standing: Use monroya.ai to see your current Share of Voice.
2. Identify missing attributes: See what the AI says about competitors that it doesn't say about you (e.g., "easiest to use," "best for enterprise").
3. Deploy structured content: Update your site with clear, declarative statements that LLMs can easily parse.
4. Seed community platforms: Increase activity on Reddit and niche forums, as these are primary sources for Perplexity and ChatGPT.
5. Monitor and iterate: Re-scan every 7 days to see how the model's perception of your brand shifts.

For teams deciding on a platform, comparing otterly.ai vs monroya.ai is common. While Otterly is often built for agencies, monroya.ai is designed for B2B SaaS teams who need to monitor the full Buyer Journey and take immediate action. You can find detailed plans on the monroya.ai/pricing page, starting with the Intelligence tier at $299/month for full category tracking.

Tracking your visibility is no longer optional. If you aren't monitoring how these models describe your pricing, comparing AthenaHQ vs monroya.ai or OmniSEO vs monroya.ai, you are leaving your brand reputation to chance. The models are already talking about you; the only question is whether you are influencing the conversation.

Find out where AI ranks you, then fix it.

FAQ

How do I measure my brand's AI share of voice?

Pick the questions your buyers actually ask, run each one repeatedly across ChatGPT, Claude, Gemini, and Perplexity on a fixed cadence, and record the share of answers that name your brand. Score how you appear, not only whether you appear, and log the sources each answer cited so you know which pages earned the mention.

How many runs are enough to trust the number?

Enough that a single unusual answer cannot move the result. Model output varies between identical prompts, so one run of one question is an anecdote. In practice you want each question asked many times on each assistant, repeated on a regular cadence, and you judge movement on the trend across scans rather than on any single reading.

What counts as a mention?

Three things are worth separating: being recommended by name, being listed among several options, and being cited as a source without being recommended. They are not equivalent, and a measurement that adds them together will overstate your position. Score them separately, then weight them.

How is this different from traditional share of voice?

Traditional share of voice counts your presence in media coverage, advertising, or search results against competitors. AI share of voice counts your presence inside the answers assistants give, where there is often only room for three or four vendors and the buyer may never click anything. The pool is smaller and the winner-take-most effect is stronger.

What is a good AI share of voice score?

It depends entirely on how crowded the category is. In a category with four serious vendors, being named in a quarter of answers is par and anything above that is a lead. In a category with forty, a much smaller share can still be leading. The useful comparison is against the named competitors in your own category, and against your own previous scans.

How long before it moves?

Assistants that read the live web, such as Perplexity and Google AI Overviews, can reflect a new or corrected page within days. Facts held inside a model's trained knowledge change only when that model refreshes, which is measured in months. Expect early movement on the live-retrieval assistants and judge the whole picture quarterly.

Check your own AI visibility

Run a free AI visibility check at monroya.ai/check — no signup, no credit card. You will see how ChatGPT, Claude, Gemini, and Perplexity describe your category and whether your brand gets named.

Monroya starts at $79/mo (Starter), with Growth at $199/mo and Intelligence at $299/mo. Every plan includes a 7-day free trial.