# AI Share of Voice

> AI share of voice is the proportion of AI-generated answers for a set of prompts that mention a brand, measured against its competitors.

Source: https://metavert.io/ai-share-of-voice  
Published: 2026-10-07  
Updated: 2026-10-07

**AI share of voice** is the proportion of AI-generated answers, across a defined set of prompts, in which a brand is mentioned or recommended, expressed relative to its competitors. It adapts an old advertising metric to [AI search](https://metavert.io/ai-search), where the question is no longer where a page ranks but whether the brand appears in the answer at all. It is the headline number in most AI visibility dashboards, and it is also the easiest one to misread, because a single run of a prompt says very little.

### How It Is Computed

The calculation has four ingredients: a set of prompts that represent how buyers ask about a category, a set of engines, a number of repeated runs, and a rule for what counts as an appearance. The simplest version divides the answers that mention the brand by all answers collected. Common variants divide a brand's mentions by all brand mentions in the category, count only answers where the brand is recommended, or weight by the order in which brands are named.

Share of voice counts brand mentions in the answer text. It is distinct from [AI citations](https://metavert.io/ai-citations), which count the linked sources. A brand can be recommended in an answer that cites only third-party pages, and a brand's page can be cited in an answer that recommends a competitor. Because the choice of prompts and counting rule drives the result, numbers from two tools, or two prompt sets, are not comparable unless those choices match.

### Why Single Runs Mislead

AI answers are sampled, not looked up, so the same prompt returns different brands on different days. Schulte, Bleeker and Kaufmann (arXiv, April 2026) measured day-to-day Jaccard similarity of 0.45 to 0.59 for the brands named in answers. That is more stable than the cited sources (0.34 to 0.42) but still far from fixed, and the authors argue that visibility should be characterised “as a distribution rather than a single-point outcome.” Sielinski (arXiv, March 2026) reached a related conclusion about citation counts: the distributions are power-law, and “many apparent differences between domains fall within the noise floor of the measurement process.”

The inputs matter as much as the repeats. Żatuchin (arXiv, July 2026; a single-author study) found that query language explains 26.5% of variance and that adding languages and models reduces error more than re-running the same prompt. A two-point move in share of voice between two weekly snapshots may be nothing more than sampling noise. The methods that separate signal from noise are covered under [AI visibility measurement](https://metavert.io/ai-visibility-measurement), and they are what tools such as [LLM Optimizer](https://metavert.io/llmopt) exist to automate.

### Category Ownership

The largest public dataset on who leads a category comes from Semrush and Growth Memo (July 2026), covering 1,094 categories and 50,000 brands in [ChatGPT](https://metavert.io/chatgpt). Only 15.2% of categories had a clear owner, and 53.7% were unsettled. Where a clear owner existed, it kept first place 90.4% of the time month over month. Leadership is rare, and sticky once established.

Brand size shapes the starting point. Kumar (arXiv, June 2026; a vendor-authored study of 149,912 citations across five engines) reports day-one visibility of 73% for household brands, 44% for mid-market brands and 11% for niche brands. The signals that correlate with visibility are mostly off-site. In Ahrefs' study of 75,000 brands (December 2025), YouTube mentions correlated with visibility at about 0.74 in ChatGPT and AI Mode and branded web mentions at 0.66 to 0.71, while backlinks sat near 0.25 to 0.28. These are correlations, not demonstrated causes, and large brands tend to score well on all of them at once. That pattern is why share of voice is usually discussed alongside [earned media](https://metavert.io/earned-media).

### Divergence Across Engines

There is no single AI share of voice. Grossman et al. (SIGIR 2026) found cross-platform source overlap below 0.2 on the Jaccard index across 11,500 real queries, and Ahrefs' most-cited-domain rankings differ by engine: [YouTube](https://metavert.io/youtube) leads in Google AI Overviews, [Reddit](https://metavert.io/reddit) in ChatGPT and [Perplexity](https://metavert.io/perplexity) (September 2026). Different sources produce different recommendations. The same holds for coding agents: a vendor study by Armature (September 2026) found that Claude Code, Codex and Cursor agreed on the same tool only 42% of the time. A blended figure hides these gaps, so share of voice is best reported per engine.

The metric matters commercially even when it produces few clicks. Similarweb data reported by Search Engine Journal (June 2026; US desktop, consumer sectors) found people shown a brand in a ChatGPT recommendation were 2.5 times more likely to visit its site within seven days. In G2's April 2026 survey of 1,076 B2B software decision-makers, 69% said they chose a different vendor than planned based on AI guidance.

## Related Topics

- [AI Visibility Measurement](https://metavert.io/ai-visibility-measurement) — The sampling methods that make this metric trustworthy
- [AI Citations](https://metavert.io/ai-citations) — The source-link metric that sits beside brand mentions
- [Earned Media](https://metavert.io/earned-media) — The third-party coverage that correlates with mentions
- [LLM Optimizer](https://metavert.io/llmopt) — Tracks brand mention rates across AI engines
- [Generative Engine Optimization](https://metavert.io/generative-engine-optimization) — The discipline this metric scores
- [AI Search](https://metavert.io/ai-search) — The engines being measured
- [Training Data Frequency](https://metavert.io/training-data-frequency) — Why well-known brands start ahead
- [AI Referral Traffic](https://metavert.io/ai-referral-traffic) — The click-based view of the same channel

## Further Reading

- [Schulte, Bleeker & Kaufmann: Don't Measure Once](https://arxiv.org/abs/2604.07585) — arXiv, April 2026
- [Sielinski: power-law citation distributions and the noise floor](https://arxiv.org/abs/2603.08924) — arXiv, March 2026
- [Żatuchin: language and model effects on AI visibility measurement](https://arxiv.org/abs/2607.13304) — arXiv, July 2026
- [ChatGPT topic authority study](https://www.semrush.com/blog/chatgpt-topic-authority-study/) — Semrush and Growth Memo, July 2026
- [Kumar: citations and day-one visibility across five engines](https://arxiv.org/html/2606.20065) — arXiv, June 2026
- [AI brand visibility correlations across 75,000 brands](https://ahrefs.com/blog/ai-brand-visibility-correlations) — Ahrefs, December 2025
- [Grossman et al.: cross-platform source overlap in AI search](https://arxiv.org/abs/2604.27790) — arXiv / SIGIR, 2026
- [Most-cited domains in AI Overviews](https://ahrefs.com/blog/most-cited-domains-ai-overviews/) — Ahrefs, September 2026
- [Most-cited domains in ChatGPT](https://ahrefs.com/blog/most-cited-domains-in-chatgpt/) — Ahrefs, September 2026
- [Most-cited domains in Perplexity](https://ahrefs.com/blog/most-cited-domains-perplexity/) — Ahrefs, September 2026
- [Which tools coding agents install](https://armature.tech/blog/which-tools-coding-agents-install) — Armature, September 2026
- [AI-recommended brands saw 2.5x more site visits](https://www.searchenginejournal.com/ai-recommended-brands-saw-2-5x-more-site-visits-similarweb/580241/) — Search Engine Journal, June 2026
- [Half of B2B software buyers now start their research with AI chatbots](https://www.prnewswire.com/news-releases/new-g2-research-half-of-b2b-software-buyers-now-start-their-research-with-ai-chatbots-302742807.html) — G2, April 2026