Answer visibility
AI Search Visibility Scorecard
Turn a fixed prompt sample into mention, citation, factual-accuracy and competitor-gap signals. Keep the model, market, prompts and date attached.
Use the same prompt set, market, model, retrieval state and date cadence when you retest. A one-off answer is a sample, not a permanent ranking.
A mention is not a citation. A citation is not a recommendation. The scorecard keeps those events separate and shows whether the sampled competitor leads or trails.
How to use it
- A mention is not a citation. A citation is not a recommendation.
- Record the model, retrieval state, location, date and exact prompt set.
- Treat model answers as volatile samples, not permanent rankings.
Score visibility
Freeze the prompt set, market, model, retrieval state and date. Record exact answers and cited URLs so the next test can show what actually changed.
Nothing is stored or submitted. The calculation runs in your browser.
Methodology
- Mention rate = prompts where your brand appears ÷ prompts tested.
- Citation rate = prompts linking or attributing to your site ÷ prompts tested.
- Fact accuracy = factually correct brand mentions ÷ prompts where your brand appears.
- Competitor gap = competitor mention rate − your mention rate. A positive result means the competitor leads in this sample.
Trace the score back to pages and public evidence.
We can inspect the prompts, competing brands and cited sources, then build the website and authority changes the scorecard alone cannot deliver.