AI Health
Monitor the operational health, performance, input quality and stability of AI capabilities across the platform.
Production Models
5
Healthy Models
2
Requiring Attention
3
Reviewed as AI signals, not incidents
Average Accuracy
95.1%
Project target ≥ 95%
Active AI Signals
6
Accuracy is compared to the project reference of ≥ 95% only where a model uses accuracy as its primary evaluation metric. Other capabilities are monitored against their own primary metric.
AI Capability Health
Platform-wide operational health of AI capabilities. Detailed model governance remains in AI & Models → Model Monitoring.
AI Performance Over Time
Only metrics relevant to the selected capability are shown.
Model Stability
Prototype monitoring statuses. No contractual drift threshold is defined.
AI Input Health
Freshness, completeness and schema health of datasets feeding production AI capabilities.
AI Operational Signals
Signals are routed to Operations for human review. Retraining is never triggered automatically.
- Platform / Data / AI / Integration
- Continuous Monitoring
- Health Signal
- Threshold or Abnormal Behaviour
- Alert
- Operations — Human Review
Prototype monitoring data. Health values, resource consumption, error rates and log events are simulated for user-experience demonstration. Health status is not an incident priority; confirmed problems are assessed in Operations.