Share of voice in AI answers

The Citead team, updated

In short. Share of voice in AI answers expresses a brand’s share of measured visibility against a defined comparison set, using a stated counting or weighting method.

Definition: Share of voice in AI answers is a brand's proportion of measured visibility relative to other brands in a defined set of AI responses. The calculation depends on the reporting method, including whether it counts mentions or weights their potential exposure.

Key facts

As of 17 October 2026, the following product definitions appear in Semrush and Ahrefs documentation.

  • Semrush's AI Visibility Metrics defines share of voice as the percentage of brand mentions relative to competitors in a market.
  • Ahrefs' AI Visibility Metrics defines AI share of voice as a brand's share of impressions among tracked brands.
  • In the same Ahrefs documentation, impressions use Google search volumes associated with prompts, and combined platform results use an impression-weighted average.
  • Semrush's metrics documentation records mentions and domain citations separately. A mention records a brand's inclusion in an answer; a citation records use of its domain as a source.

How is a mention-based share calculated?

Semrush's measurement guide gives a formula that divides a brand's mentions by all brand mentions in its category. Expressed as a calculation, it is:

Share of voice (%) = (brand mentions / total category-brand mentions) * 100

The guide's illustrative example uses 10 brand mentions out of 100 total mentions, producing a share of 10%. The denominator is the total count of brand mentions included in the comparison. An answer containing several brands can contribute to several brands' counts.

A prompt is an input submitted to an AI assistant. Ahrefs' counting rule counts a brand once when it appears at least once in a response. Recording that rule makes the unit of measurement explicit: repeated appearances of the same name within an answer receive a single count.

Why can tools report different percentages?

The mention-based formula is one calculation described in Semrush's guide. The same guide says its Brand Performance report also considers how high a brand appears in an answer. A dashboard percentage therefore needs its product-specific methodology alongside it.

Ahrefs' impression-based definition introduces a different weighting: prompts associated with greater Google search volume contribute more potential exposure. Its combined-platform calculation also gives greater weight to platforms with more impressions. These percentages describe different measured quantities, so comparisons require matching definitions and brand sets.

What should accompany a report?

A useful report records its comparison scope. Semrush's Brand Performance documentation specifies a target location and language at setup, a competitor selection, and an AI platform filter. It also provides a selector for weekly historical snapshots. These settings belong beside the percentage when comparing reporting periods.

The same documentation describes an Answers for panel containing full responses and an export of the Narrative Drivers table to CSV or Excel. Reviewing the underlying answers lets a website owner check which brands appear and which domains supply cited material. Keeping mention totals beside share of voice also helps explain a change caused by the comparison denominator.

FAQ

Does share of voice measure website citations?

Check the metric's definition. Semrush's metrics page separates brand mentions from responses citing a domain, so a mention-based share and a citation count measure different events.

Can my share change if my mention count stays constant?

Yes. In Semrush's mention-based formula, your share falls when the total category-brand mentions increase while your own count stays constant. Keep the competitor set and calculation method consistent when interpreting changes.

Sources

  1. Semrush, "AI Visibility Metrics", 22 September 2025
  2. Ahrefs, "AI Visibility Metrics", 3 September 2026
  3. Semrush, "How to measure AI share of voice using Semrush", updated 17 July 2026
  4. Semrush, "AI Visibility Brand Performance Reports", 23 September 2025

Check how a page reads to answer engines. Open the GEO score analyzer

Test a page and see what to fix first

The analysis usually takes less than 10 seconds and needs no account.

400,000+ pages scored in the public benchmark