# AI pilot case register

Workflow/version: [assign] | Case-set owner: [assign] | Reviewed date: [date or unknown]

Write expected behavior before running the case. Actual results stay blank until observed. Preserve failures and changed expectations; do not overwrite a failed run with a later passing result.

| Case ID | Category | Fictional/approved input or source reference | Expected behavior | Observable pass criterion | Critical stop condition | Actual result + artifact reference | Reviewer/date | Status |
|---|---|---|---|---|---|---|---|---|
| [ID] | Ordinary | [input/reference] | [behavior] | [criterion] | [condition/to agree] | [not run] | [unassigned] | Draft |
| [ID] | Missing information | [input/reference] | [behavior] | [criterion] | [condition/to agree] | [not run] | [unassigned] | Draft |
| [ID] | Conflicting instructions | [input/reference] | [behavior] | [criterion] | [condition/to agree] | [not run] | [unassigned] | Draft |
| [ID] | Follow-up correction | [input/reference] | [behavior] | [criterion] | [condition/to agree] | [not run] | [unassigned] | Draft |
| [ID] | Workflow-specific exception | [input/reference] | [behavior] | [criterion] | [condition/to agree] | [not run] | [unassigned] | Draft |

## Coverage statement

- Why these cases were selected:
- Actual source population and dates:
- Cases/conditions excluded:
- Differences between this exercise and the intended live workflow:
- Further cases needed and owner:
- What this case set can support:
- What remains untested:

## Retest history

| Run ID | Case ID | System/config version | Date | Observed result | Exception ID | Reviewer | Next step |
|---|---|---|---|---|---|---|---|
| [ID] | [ID] | [version] | [date] | [actual result] | [if any] | [name] | [action/wait] |

A small case set can clarify a design. It does not by itself establish performance across the business or future inputs.