Skip to main content
Competitor share-of-voice shows how often each brand — yours and your competitors’ — is named across your tracked prompt set, expressed as a proportion of all brand mentions. It answers: when a buyer asks an AI assistant about this category, whose name comes up?

​
Where to see it

Open Analytics → AI Visibility:
  • The AI Share of Voice card shows your share of all brand mentions in the latest answers and your rank among the brands named.
  • Rankings lists every brand the assistants named across your tracked prompts.
  • Visibility by Assistant shows how often each brand is named on ChatGPT, Claude, Gemini and Perplexity.
  • The Prompts table shows the top brand for each prompt.
See AI Visibility.

​
How the competitive set is built

Ryze records every brand named in the answers — there is no competitor list to maintain — so the resulting set is frequently different from the one you track commercially. Two things usually surface:
  • Established players you do not compete with directly but who dominate category-level questions, because they are strongly represented in training data.
  • Adjacent-category tools the model considers substitutable even when buyers do not.
Both matter, because they are what a prospect is being shown.

​
Reading the number

Compare prompt by prompt, not only the blended figure. The Prompts table shows which brand leads each question: a brand can dominate recommendation-style prompts and be absent from problem-led ones, and the average conceals exactly the structure you would act on. Phrase your prompts by intent when you add them so the table is easy to read. Look at the gap, not your absolute figure. Going from 2% to 4% while the leader holds 30% is real progress but not a competitive change. Separate mentions from citations. High share of voice with low citation share means competitors are named from memory rather than from anything the model read — a different problem with a much slower remedy.

​
What moves it, and what does not

Mentions come from the model’s training data and move slowly; citations come from live retrieval and respond to content — see AI visibility.
  • Moves it slowly, over months to years: sustained third-party presence — reviews, comparison sites, forums, press, and being written about by people who are not you.
  • Does not move it: publishing more of your own pages. Your own content can earn citations, but it rarely makes a model name you from memory.
  • Moves it indirectly: being cited consistently over time contributes to broader presence. This is a long, indirect path, not a tactic.

​
Setting expectations

If you are starting from zero mentions, the honest sequence is: citations first (weeks to months), mentions much later (many months), competitive share last. Content work buys the first. Only sustained presence buys the rest. Treat any claim that share of voice can be moved quickly with suspicion.