AI share of voice is how often AI assistants mention your brand versus a competitor on the buying questions your category gets asked. Run a free head-to-head: enter your website and one competitor, and get a report card showing which brand AI names in each answer — and who wins overall.
Give us your website, one competitor’s website, and your work email. Magister reads both sites to identify each brand and the category you compete in — that context is what makes the questions specific to your market.
It generates about five real buying questions for your category — deliberately unbranded, the way an actual prospect asks — because a question that already names either brand can’t measure share of voice.
The questions run live against Claude, and every answer is scored the same way: is your brand mentioned, is the competitor mentioned, and in what position each appears.
Your report lands in your inbox as a shareable link: the verdict, a two-bar comparison, and the question-by-question breakdown of who got named where. Built to screenshot.
One honest headline: which brand AI assistants mention in more answers, with the exact counts — or a dead heat when neither side wins.
Your mentions versus your competitor’s mentions across the full question set, side by side — the screenshot that says it all.
Every buying question with whether each brand was named in the answer and its position when it was — so you see exactly which questions you win and which you lose.
The report lives at a private link with a social preview card, so you can drop it in Slack or a client thread and it unfurls with the verdict.
AI share of voice is the portion of AI-assistant answers that mention your brand versus competitors when people ask buying questions in your category. It’s the AI-answer equivalent of search share of voice: as more product research ends inside ChatGPT-style answers instead of a page of links, being the brand the answer names is the visibility that matters.
This free check runs about 5 unbranded category questions against Claude and scores each answer for both brands. That makes it a snapshot — one assistant, one question set, one moment in time — not a longitudinal tracker. The full Magister product tracks 5 engines (ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews) over time, so you get a trend instead of a single reading.
A question that already names a brand can’t measure share of voice — of course the answer mentions the brand you asked about. Unbranded questions ("best X for Y", "how do I solve Z") are how real prospects ask before they know who the players are, so they’re the answers worth winning.
Yes — no credit card, no account. The head-to-head runs in the background and takes a few minutes, so the report is delivered to your inbox as a private shareable link rather than making you wait on the page.
That’s the useful outcome — now you know, question by question, where AI recommends your competitor instead of you. AI answers draw on sources like reviews, comparison pages, and community threads; on paid plans the Magister agent does the work of earning those citations and tracks whether your share of voice moves.
This check is one signal. The free audit grades your whole site for AI search and shows the plan Magister would run to fix what it finds.