AI share of voice

How do you calculate AI share of voice?#

Calculate AI share of voice by running a fixed set of buyer prompts, counting the answers that name or cite each brand, and dividing your count by the total — either by all answers, or by all brand appearances if you want your slice of the whole conversation.

There are two common versions, and they answer different questions. Vendors and agencies use the term loosely, so report both and say which one you mean.

Count mentions and citations separately as well. A mention is your name in the answer text; an AI citation is a link to your page. The engines differ sharply: in Semrush’s June 2026 study, 87% of the times a brand’s domain appeared in a ChatGPT answer it was a linked citation, but only 20.7% named the brand in the text. Gemini was the reverse, at 21.4% linked and 83.7% named. One blended number hides which kind of visibility you are winning.

The raw counts come from prompt tracking: the same buyer questions, run on a schedule, with every brand and cited page recorded.

Why does AI share of voice change from week to week?#

AI share of voice changes because AI answers are probabilistic: the same prompt returns a different set of brands on almost every run, so a single check is noise and only a large, repeated sample shows a real trend.

The SparkToro study — 600 volunteers running 12 prompts through ChatGPT, Claude and Google’s AI 2,961 times — also found the signal inside the noise: how often a brand appears across many runs is stable enough to track. For one Google AI prompt about digital marketing consultants, a single agency appeared in 85 of 95 responses. Frequency across runs, not position in any one list, is what share of voice should be built on.

What is a good AI share of voice?#

A good AI share of voice is one that beats the competitors you meet in deals and keeps rising; there is no published universal benchmark, because the number depends entirely on which prompts and which rivals you chose to track.

Two brands can report 40% and 12% in the same market and both be right: one tracked ten category prompts against two rivals, the other fifty prompts against eight. Share of voice is only comparable inside one fixed panel. What makes it meaningful:

• Rank against named rivals. Leading your own panel matters more than any absolute figure.
• Split by prompt type. Category prompts (“best CRM for startups”) and problem prompts (“how to stop losing leads”) usually tell different stories; a brand can own one and be missing from the other.
• Split by engine. Engines search different indexes. Claude and ChatGPT shared only about 8% of their cited domains in Profound’s 2026 data, so a strong ChatGPT share says little about Claude.
• Watch the trend. A share that climbs from 10% to 18% over a quarter is a better result than a flat 25%.

How do you improve AI share of voice?#

Improve AI share of voice by owning the specific questions where competitors are named and you aren’t: publish a clear, answer-first page for each one, and earn mentions on the sites the engines already cite for your category.

1. Find the gaps in your panel. List the prompts where a rival appears and you don’t, and note which pages the engine cited instead.
2. Answer the sub-questions. Engines fan each prompt out into narrower searches, so a page that wins one sub-query can be cited even when you don’t rank for the head term.
3. Build official pages for facts. Pricing, integrations and comparison pages on your own domain are what engines check before they recommend you.
4. Earn off-site mentions. In Ahrefs’ study of 75,000 brands in Google AI Overviews, branded web mentions correlated with AI visibility at 0.664, against 0.326 for Domain Rating and 0.218 for backlinks. See brand mentions.
5. Re-run the same panel. Compare month over month, so you can see which change moved the share.

Each engine has its own route in — different crawlers, indexes and citation habits. The AI SEO guides compare them side by side.

Common mistakes with AI share of voice#

The most common AI share-of-voice mistakes are weighting brands by list position, counting prompts that contain your own name, and comparing numbers from different prompt panels — each one moves the metric for reasons that have nothing to do with buyers.

Related terms#

• MeasurementPrompt trackingThe practice of running a fixed set of buyer questions through AI assistants such as ChatGPT, Perplexity and Gemini on a schedule and recording which brands and pages each answer mentions or cites — the AI-search equivalent of rank tracking.Read the entry
• AI search & GEOAI visibilityA measure of how often, how prominently and how favorably a brand appears in answers from AI systems such as ChatGPT, Perplexity, Gemini and Google AI Overviews for the questions its buyers ask — the AI-search counterpart to search rankings.Read the entry
• AI search & GEOAI citationA link or source reference that an AI answer engine attaches to its response to show where a claim came from, and it differs from a brand mention in that it points to a specific page that can be clicked and can send traffic.Read the entry
• Authority & off-siteBrand mentionsAny reference to a brand’s name on another website, forum, review site or video, with or without a link, and in large studies mentions correlate more strongly with visibility in AI answers than backlinks do.Read the entry
• GEOAI search & GEOGenerative engine optimizationThe practice of making content easy for AI answer engines such as ChatGPT, Perplexity and Google AI Overviews to retrieve, quote and cite — where SEO competes for a ranked link, GEO competes to be a named source inside the generated answer.Read the entry
• MeasurementAI referral trafficThe visits a website receives from links inside AI assistants and answer engines such as ChatGPT, Perplexity, Gemini, Claude and Copilot, identified in analytics by referrer domains like chatgpt.com and perplexity.ai.Read the entry