The Incident Management feature in RexCommand is used to capture, document, and manage unexpected or adverse events that occur in connection with AI systems. These incidents may include hallucinations, system failures, ethical violations, privacy breaches, or any other high-impact issue that requires review, escalation, or resolution.
Purpose
Incident reporting ensures your organization can respond quickly and consistently to problems that may compromise system performance, safety, trust, or compliance.
Incidents may originate from internal monitoring, audits, or public reporting mechanisms such as the external feedback form.
What You Can Track
Each incident entry includes:
Title and Description – Short summary of what happened
Status – Current state (e.g., Detected, In Review, Resolved)
Severity – Impact level (e.g., Low, Medium, High, Critical)
Detected Timestamp – When the incident was reported or discovered
Actions – Access to View Details for full incident context
Key Features
+ Report Incident – Submit a new incident to log a significant event. Learn how to report an incident
Search and Filter – Use filters to find incidents by status or severity, or search by keyword
View Details – Review full incident history, linked systems, response status, and comments. Learn more about viewing incident details
Examples of Incidents
Data breach or unauthorized access
AI-generated hallucinations or unethical responses
System outages, failed backups, or monitoring failures
Model behaving inconsistently or outside defined safety bounds
Use of unapproved datasets or drift in sensitive models
Notes
Use this dashboard to maintain a log of all known incidents across your AI and data estate
Keeping this record up to date helps ensure transparency, accountability, and readiness for audit or regulatory response
Incident response workflows are typically handled in collaboration with risk, compliance, engineering, and legal teams