Share of voice shows how often you appear in answers compared with the competitors you name. We split it by assistant and by topic, because a strong showing on one model can hide a weak one elsewhere. The figure is a measure of what was observed on the questions you track, not a claim about the whole market.
Choosing the competitor set matters. Too narrow and the number looks flattering. Too wide and it becomes noise. We agree the list with you, including the rivals you rarely think about but customers keep asking about, and revisit it when the market changes. A quarterly check keeps the comparison honest.
The weekly trend report is short on purpose. It names the movements worth a conversation and leaves the raw answers one click away for anyone who wants to check the detail. A monthly review goes further and ties movement to the work we did.
A change in share of voice can come from your work, a competitor's work or a shift in the assistant itself. The agent flags movements and lists likely causes, such as a new article or a changed listing, but it labels them as possibilities. We do not claim a cause unless the evidence is clear.
This agent suits teams with at least a couple of known competitors. If you operate in a niche where assistants name almost no one, a share figure can swing on very few answers, and we would tell you to read it with extra caution.
ChatGPT, Gemini, Perplexity, Claude and Copilot, with more added as they matter.