Churn propensity model accuracy below approved target

Model monitoring reported accuracy below the approved target for two consecutive evaluation windows, affecting predictive outputs used in CX prioritisation.

INC-1039HighPending ResolutionPRODOwner: AI Analyst 03
Not ApplicableTarget: To Be Confirmed

Impact

Medium

Urgency

High

Priority

High

Created

17 Aug 2026, 10:02

SLA Target

To Be Confirmed

Configured according to approved SLA once formally defined.

Time Remaining

Target: To Be Confirmed

  1. Created
  2. Assigned
  3. Investigating
  4. Resolution Applied
  5. Verification
  6. Resolved
  7. Closed

Incident overview

Description
Model monitoring reported accuracy below the approved target for two consecutive evaluation windows, affecting predictive outputs used in CX prioritisation.
Category
AI Model
Affected area
Predictive customer analytics
Environment
PROD
Impact
Medium
Urgency
High
Priority
High
Assigned team
AI Operations Team
Owner
AI Analyst 03
Source
Monitoring
Last update
17 Aug 2026, 12:10
Priority override
None — derived from Impact + Urgency

Customer / business impact

Affected capability
Predictive customer analytics
Potential effect
Churn prioritisation lists may be less reliable until the model is validated.
Affected users / scope
CX prioritisation users (prototype scope)
Business impact
Retention actions may be targeted less precisely during the affected period.

Priority assessment

Impact × Urgency matrix (prototype logic).

Impact ↓ / Urgency →HighMediumLow
High
Medium
Low

Prototype operational logic. Priority is derived from Impact + Urgency and is not presented as a contractually approved mapping.

Prototype operational data. Values, references and identities are simulated and do not represent live NWC operations.