N NoBullshit.wiki 中文

Full-service restaurant field supervisor rectification audit - findings from consecutive store visits closed out one by one, separating "fixed", "fixed on paper only", and "fixed then relapsed", with the repeat-offender list ranked by recurrence count

Works with:Claude / GPT-4 classCategories:Full-Service RestaurantLevel:Advanced Updated

You are a field supervisor coach for a full-service restaurant chain. I will give you the findings from the last [N] store visits for one location plus this visit's on-site verification. Run the rectification closure audit.

Inputs:
- Findings table from each past visit: date / finding / area (front of house, kitchen, storage, cashier) / manager's sign-off at the time
- This visit's verification: the on-floor status of each finding (transcribe the photos if that is what you have)
- Any rectification write-up the manager submitted in between

Closure standard (four fixed categories; physical evidence only - no "already fixed" statement counts):
1. Closed: a visible fix on the floor, not relapsed since the last verification
2. Paper-closed: the manager's write-up says fixed, verification finds no trace of it - count as NOT fixed
3. Relapsed: previously verified closed, back on the floor this visit
4. New: first appearance

Output format:
- Closure ledger: finding / first-seen date / times appeared / status this visit (one of the four) / area
- Repeat-offender list: ranked by consecutive appearances, the top three itemized
- Paper-closed list: each item showing the gap between the manager's statement and the floor reality
- Three closing actions only: the first must be the escalation recommendation for the #1 repeat offender (one more recurrence converts to a formal talk); the other two ranked by impact on food output and guests

Constraints:
- Past findings not covered by this visit's verification get "not verified this visit" - never default them to closed
- No commentary on the manager's attitude, no inferring causes - the audit closes out findings; any judgment about people stays out of the output
- No "strengthen execution" or "raise awareness" recommendations with no verification attached
- Keep the original wording of each finding across reports - do not merge similar items for me; the merging call is mine, not yours

Why it is written this way

A supervisor's time mostly goes to repeat problems: find it, mention it, manager agrees, next month it is back. This audit treats "rectification" as an object with a lifecycle - found, promised, verified, closed or relapsed. Without that chain, "it was fixed" is just a sentence, equally worthless in whichever report it appears.

The sharpest of the four categories is paper-closed: the manager's statement versus the floor reality, listed side by side, separating "said it was fixed" from "it was fixed" at the statistics level. Without the split, the closure rate always looks good, the floor always looks bad, and the supervisor's monthly report is a distorted number in front of the regional manager.

The repeat-offender list ranks by consecutive appearances rather than severity because severity invites debate (what you consider important, the manager considers trivial) while counts do not: a finding appearing in its third report, whatever its size, proves the previous two responses failed. Impact ranking is reserved for the other two actions - escalation reads the recurrence count, fixes read the impact.

The "no judgments about people" constraint is written for the data itself: the moment an audit starts rating the manager, the store starts "preparing" before the next visit - problems tucked away, let back out after the supervisor leaves. Closure judged only against on-floor traces is what lets this ledger keep running month after month.

Variables

Variable Notes
Visit cadence Weeks between visits - defines what "relapsed" means in time; write it at the top of the audit input
Verification coverage Which past findings this visit's walk-through covered - uncovered items are marked not-verified, never guessed
Escalation trigger How many consecutive appearances convert to a formal talk - align with the trigger in the escalation-talk script card

store visitsrectification trackingclosure audit