AI share of voice is the percentage of AI answers in your category that name your brand, measured against the competitors named instead. In a medium with no rankings and no results page, share of voice is the scoreboard.
Take a fixed set of buyer questions. Run each through ChatGPT, Gemini, Claude, and Perplexity multiple times. AI share of voice = your mentions divided by total brand mentions across all sampled answers. Track it per question, per assistant, per month, and it behaves like any other competitive metric: it trends, it responds to work, it predicts pipeline.
Raw mention counts flatter you in growing categories: everyone’s mentions rise as usage grows. Share of voice is zero-sum, so it shows who is actually winning the answers. A rising count with falling share means competitors are growing faster inside the same conversations, which is exactly the signal a count alone hides.
Being named first in an answer is not the same as being the “also consider” at the end. Sophisticated share-of-voice scoring weights first mentions and recommendation framing above passing references. Aethon scores both, because buyers overwhelmingly act on the first confident name.
The same levers that create AI visibility, weighted competitively: presence in the citations assistants lean on, quotable comparison content, review consensus, and coverage of the conversational moments where your competitors are currently the default answer. Every point of share you gain is a point someone else loses.
One table: rows are your priority questions, columns are the four assistants, cells are your share versus the leader’s. Executives grasp it instantly because it looks like market share, and it is: market share of the answers buyers hear. Aethon reports it monthly with the before-and-after answers that explain each move.
From your manual sampling: count answers where each brand was named, then compute your share.
The percentage of brand mentions you capture across sampled AI answers in your category, measured against competitors, per assistant and over time.
Your mentions divided by total brand mentions across repeated samples of a fixed buyer-question set run through ChatGPT, Gemini, Claude, and Perplexity.
Because AI answers are zero-sum: only a few brands get named per answer. Share shows competitive position; counts can rise while you lose ground.
Category leaders typically capture 30% or more on their core questions. The practical goal is beating your nearest competitor question by question.
Continuous sampling across the four assistants with position-weighted scoring, competitor tracking, and citation mapping that explains every shift.