# Agent Ownership Record

One record per production AI agent. A name in an owner field is not an ownership model: the record has to say what the owner is authorized and expected to decide, what pauses for a human, and what happens when the owner is gone.

Fill one in before the agent reaches production, store it with the agent inventory, and review it on the cadence set by the agent risk tier.

## Template

```text
Agent name:
Agent ID:
Business purpose:
Business workflow:
Accountable business owner:
Owner role and authority:
Technical owner:
Risk or policy owner:
Primary human reviewer:
Backup reviewer:
Escalation owner:
Systems and tools accessed:
Data categories accessed:
Actions the agent may perform:
Actions requiring human approval:
Prohibited actions:
Financial or operational thresholds:
Kill, pause, or restrict mechanism:
Rollback or compensating action:
Incident contact:
Success metric:
Error or intervention metric:
Evidence and log location:
Last review date:
Next review date:
Owner succession rule:
Retirement criteria:
```

## Field notes

- **Accountable business owner**: one named individual who owns the outcome the agent can change, not a team alias or a committee. If the field says "AI team" or "IT", nobody is accountable.
- **Owner role and authority**: state what this person can approve, fund, restrict, and stop. Responsibility without authority is the most common failure mode.
- **Technical owner**: agent identity, integrations, permissions, reliability, monitoring, rollback, and shutdown.
- **Risk or policy owner**: security, privacy, legal classification, evidence requirements, and control standards.
- **Primary and backup reviewer**: the humans who decide the actions the agent may not decide alone. A single reviewer with no backup is a single point of failure.
- **Escalation owner**: who receives expired approvals, unresolved exceptions, and incidents, and who can stop the workflow.
- **Actions requiring human approval**: separate read, recommend, draft, approve, execute, and irreversible actions. Approve or gate by action class, not by whole agent.
- **Thresholds**: the amount, volume, or blast radius above which a higher approver is required.
- **Kill, pause, or restrict mechanism**: the exact control and who can use it, tested at least once before launch.
- **Next review date**: quarterly for read-only internal agents, monthly for customer-facing agents, weekly plus continuous monitoring for agents that move money, change access, or write to production. Any model, prompt, tool, permission, or workflow change triggers an off-cycle review.
- **Owner succession rule**: who takes over if the owner changes role or leaves, and what happens if no successor exists. The default should be restrict, pause, or retire, never "keep running".

## Multi-agent workflows

Assign ownership at three layers and keep a record for each:

1. **End-to-end workflow**: one accountable business sponsor for the overall result.
2. **Material agent or sub-agent**: a named domain or technical owner for its behavior, dependencies, permissions, and delegation trace.
3. **High-impact capability**: an owner and escalation path per sensitive action and resource pair, such as issuing a discount or changing access.

Ephemeral agents that exist for seconds inherit ownership from their blueprint, policy, workflow, or approved capability class, not from the temporary runtime process.

Source: Contro1 Agent Ownership Record, 2026. Adapt the fields to your organization and record named owners in the agent inventory.