Incident Management
For high-risk AI systems, deployers must monitor operation and report serious incidents. Klarvo's incident management helps you track, respond to, and document AI-related incidents.
What Counts as an Incident?
AI incidents requiring documentation include:
Incident Severity Levels
| Level | Description | Response Time |
| Critical | Immediate harm, safety risk, prohibited practice | < 24 hours |
| High | Significant impact, rights violation | < 48 hours |
| Medium | Moderate impact, contained | < 1 week |
| Low | Minor issue, no harm | < 2 weeks |
Incident Workflow
Detection → Logging → Triage → Containment → Investigation↓ ↓
Notify Resolution
↓ ↓
Internal teams Postmortem
↓ ↓
Provider (if needed) Reassessment
↓ ↓
Authority (if serious) Closure
Required Documentation
Each incident record should capture:
Basic Information
Timeline
Impact Assessment
Response Actions
Resolution
Serious Incident Reporting
Under Article 26(5), deployers must report serious incidents to:
A "serious incident" includes:
Integration with Reassessment
Incidents can trigger system reassessment:
Best Practices
🚨 Log immediately: Don't wait to document
📞 Notify early: Err on the side of over-communication
🔍 Root cause: Always dig to underlying issues
📝 Postmortem: Learn and prevent recurrence
🔄 Update procedures: Improve based on learnings