# AI visibility prompt register

Use this worksheet to define the sample before checking it. Editable fields are intentional.

## Study scope

- Business and website: [name/URL]
- Audience and decision: [scope]
- Market and language: [settings]
- Actual surfaces/providers: [source and retrieval method]
- Schedule: [cadence]
- Brand matching method: [exact/variant rules and manual review]
- Citation matching method: [domain/subdomain/URL rules]
- Known limitations: [scope]

## Stable prompts

| ID | Exact prompt | Intent | Branded/unbranded | Why included | Active from | Retired/date/reason |
|---|---|---|---|---|---|---|
| [ID] | [question] | [intent] | [type] | [reason] | [date] | [history] |

## Run evidence

| Run ID | Date/time | Prompt ID | Actual surface/model if known | Settings | Status | Mention | Cited URLs | Recommendation context | Evidence file |
|---|---|---|---|---|---|---|---|---|---|
| [ID] | [timestamp] | [ID] | [source] | [settings] | [completed/failed/unavailable] | [yes/no/unavailable] | [URLs/unavailable] | [context] | [path] |

## Summary calculations

- Attempted checks: [A]
- Completed usable answers: [C]
- Failed/unavailable checks: [A − C]
- Answers mentioning brand: [M]
- Answers citing website: [S]
- Mention rate: M/C, undefined if C=0.
- Website citation rate: S/C, undefined if C=0.
- Completion rate: C/A, undefined if A=0.

Mention and citation sets may overlap. A model API response must not be labeled a consumer chatbot answer or Google AI Overview unless that is the surface actually measured.

## Action record

| Observation | Evidence | Factual/content change | Owner | Verification | Next comparable run |
|---|---|---|---|---|---|
| [observation] | [link] | [action] | [owner] | [check] | [date] |