What this measures (and what it does not)

RhinocerosAI answers one question repeatedly and carefully: when someone asks an AI engine a question in your client's category, what does the engine say, and does your client appear in the answer? Everything in the product exists to make that question answerable with evidence rather than assertion.

In scope

• Prompt tracking. A per-brand set of questions, run on a schedule across every engine in the brand's market.
• Mention detection. For each answer: was the brand named, how prominently, in what company, and with what tone.
• Share of voice. The brand's share of mentions against a declared competitor set, per engine and aggregated by the market's engine mix.
• Citation and source analysis. Which URLs and domains the engine cited, and whether the brand owned any of them.
• Trend and history. Every metric over time, with model-version changes marked on the line.
• AI-readiness diagnostics. A read-only check of whether AI crawlers can reach and parse the pages you want cited.

Deliberately out of scope

The platform does not write content, does not edit your client's site, and does not submit anything anywhere. It is the instrument, not the intervention. The reason is credibility: a tool that both scores you and sells you the fix has an incentive problem that no amount of disclosure repairs.

What a measurement actually is

One measurement is a single run: one prompt, one engine, one language, one fresh session, one moment in time. A run produces an answer, and that answer is stored verbatim. Every figure you ever see is an aggregate over runs, and the product will always tell you how many runs are behind it.

This is why nothing in the interface shows a bare percentage. See Reading a band for what replaces it.