SolvedAgainGet your scorecard

Platform · Measure

One number your team can rally around. Five more that explain it.

The scorecard turns 300 grounded AI answers into a Share of AI Voice you can put in front of anyone — and every number on it traces back to a stored answer you can open and read.

AI Visibility Scorecard

Grounded · live web search

Share of AI Voice

24% +6

Questions won

7/20 +3

Avg. rank

2.3

Share of AI Voice

Your brand24%
Competitor A48%
Competitor B28%

90-day trend

▲ climbing

ChatGPTClaudeGeminiPerplexityGrok

Illustrative example — your report shows your own numbers, each tracing to a stored answer.

Diagnosis · Authority gap

The AI cited a “best of” roundup you're not listed in. Fix: get into that exact roundup.

Lost → Won

“best expense platform for startups?”

Six metrics, zero black box.

Most visibility tools compress everything into a proprietary index you have to take on faith. We publish the six numbers and the formula behind each one — computed as pure, unit-tested functions over your stored extractions.

Mention rate
How often your brand comes up at all, across every sample.
Share of AI Voice
Your mentions divided by every tracked brand's mentions — your actual slice of the answer.
Average rank
Where you land among the brands named, on the samples where you show up.
Recommendation rate
How often a mention is a real endorsement, not just a name-check.
Citation share
How many of the sources the AI cited are yours.
Volatility
How much the 3 samples disagreed — shown, not hidden.
Unprompted visibility
Your numbers on questions that name no tracked brand — the way a cold buyer asks. Questions that name you or a competitor are scored separately, so they never inflate it.
Source Map
Every domain the engines cited, ranked by how many answers cited it — the sources that actually decide your category's answers, and your outreach list for getting into them.
Recognition ladder
Mention rate by question specificity, from branded questions down to fully generic ones. The deepest rung where you're still named is your recognition boundary.

Per engine, not blended

ChatGPT and Claude on the standard audit; Gemini, Perplexity and Grok join on Full Coverage and every subscription plan. Each engine gets its own breakdown, because being invisible on one engine and strong on another is a strategy question, not a rounding error.

Grounded, and labeled

Every question runs with live web search on, and every answer stores its grounded flag. An ungrounded answer measures what a model memorized in training — a different thing entirely — and the scorecard never mixes the two silently.

Movement, month over month

On subscription plans the same approved questions re-run automatically each month, so the delta compares like with like: share moved, rank moved, questions flipped from lost to won.

Read the full methodology →

Start with the free scorecard.

Your Share of AI Voice against up to 4 competitors — free, grounded, in about 20 minutes.

$0 to start · no card required · every number traces to a stored answer