Cleaning Coverage Accuracy Rate (CCAR)

GMI-FAC-017Facilities ManagementPilot-ReadyPattern O-7: Inspection Coverage

Definition

Percentage of actually-occupied spaces correctly identified as "occupied" by the AI cleaning scheduler, measured by cross-referencing against an independent occupancy signal.

The Measurement Gap

There is no metric that validates the AI cleaning scheduler's occupancy determination against independent measurement, which means demand-based cleaning may systematically skip occupied spaces that the model misclassifies as unoccupied.

Fixed cleaning schedules clean every space regardless of use. AI-optimized cleaning schedules clean based on detected occupancy: spaces classified as unoccupied are skipped or receive reduced service. The cost savings are real: cleaning labor and consumables decrease when unoccupied spaces are skipped. But the savings depend on the accuracy of the occupancy classification. A space that was occupied but misclassified as unoccupied receives no cleaning: occupants return to an unsanitary environment.

The occupancy model's classification accuracy is rarely measured against ground truth. Badge data, the most common input, has known gaps: unbadged visitors, spaces accessed through propped doors, restrooms without badge readers. A conference room that hosted a 2-hour meeting entered through an adjacent hallway may register as unoccupied in the badge system while the cleaning model skips it.

Decision Supported

Is the AI-optimized cleaning schedule accurately detecting occupied spaces, or is the model missing spaces that need cleaning because the occupancy signal has gaps the model doesn't account for?

Formula

Plain Language

Compare the cleaning scheduler's occupancy determination against an independent occupancy signal (camera motion detection, Wi-Fi client counts, manual spot-checks). Count spaces that were actually occupied but the model classified as unoccupied.

Notation

CCAR = 1 - (Occupied_but_skipped / Total_actually_occupied) x 100

Variables

Variable Definition Unit
Occupied_but_skipped Spaces classified as unoccupied by the cleaning model that the independent signal confirms were occupied Count
Total_actually_occupied All spaces confirmed as occupied by the independent signal during the measurement period Count

Units: Percentage. 100% equals perfect detection: every occupied space was correctly identified. The independent signal also has gaps, making CCAR a floor estimate.

Proposed formula, pending validation pilots.

At a Glance

Category
Service Delivery Coverage & Calibration

Decision Supported

Is the AI-optimized cleaning schedule missing spaces that were actually used?

← Visitor Risk Scoring Equity Index (VRSEI)Cleaning Intensity Calibration Score (CICS) →

Assess Your Organization

The AI-GPS Assessment evaluates your AI governance maturity and recommends KPIs like this one for your implementation roadmap.

Take the Free AI-GPS Assessment