Model / AI System Card

Overview

• System name and ID:
• Version / configuration:
• Owner:
• Intended purpose:
• Primary users:
• Deployment environment:
• Risk tier:

Intended use and limits

• In-scope tasks:
• Out-of-scope or prohibited tasks:
• Required human-review or escalation conditions:
• Known limitations and uncertainty behavior:

Components and data

• model/provider and version:
• prompt/configuration version:
• retrieval sources and update cadence:
• tools and permission scope:
• data categories, provenance, retention, and access controls:

Evaluation

Evaluation area	Method / scenario set	Result	Limitation / follow-up
Task quality
Safety / misuse
Privacy / data handling
Reliability / tool failure
Fairness / accessibility, where relevant

Monitoring and incident readiness

• monitored signals and thresholds:
• logging and evidence-retention approach:
• incident and escalation owner:
• rollback / disable path:

Change history

Date	Change	Materiality	Reviewer / decision
		minor / material / major