General

Draft an Enterprise Agent Governance Playbook

Create a comprehensive operational governance playbook for deploying automated business agents, covering risk tiers, human override, and audit trails.

How to use this prompt

Copy this prompt into your assistant, replacing the bracketed fields with your specific deployment context and organizational constraints. The assistant will reason through the operational risks before delivering a structured, production-ready playbook tailored to your requirements.

The prompt

## Role & objective
You are a principal enterprise risk architect and automation governance specialist. Your objective is to draft a rigorous, comprehensive governance playbook for deploying automated agents within a business organization, ensuring operational safety, accountability, and regulatory compliance.

## Inputs
- Business domain and use case: [e.g., customer support tier-1 resolution, invoice processing]
- Target agent architecture/capabilities: [e.g., LLM-based API caller with database write access]
- Regulatory environment: [e.g., GDPR, HIPAA, SOC 2, or none specific]
- Organization risk tolerance: [conservative, balanced, or aggressive]

## Instructions
1. Reasoning: Before drafting the playbook, briefly outline your assessment of the primary failure modes for this specific use case, state your key assumptions about the underlying tech stack, and establish the risk thresholds.
2. Risk Classification: Define a tiered risk framework (Low, Medium, High) based on autonomy level, data sensitivity, and potential financial or reputational impact.
3. Human Override: Specify exact triggers, mechanisms, and response time SLAs for human-in-the-loop intervention and emergency kill-switches.
4. Audit Trail: Outline technical logging requirements, retention periods, and review cadences to ensure complete traceability of agent decisions.
5. Deployment Readiness: Create a go/no-go gate checklist covering security testing, bias evaluation, rollback plans, and stakeholder sign-offs.
6. If any critical input is missing or ambiguous, ask 1 to 2 clarifying questions before producing the final playbook.

## Constraints
- Do not use generic corporate platitudes; specify concrete technical and operational controls.
- Maintain a formal, authoritative, and practical tone suited for chief risk officers and engineering leads.
- Quality bar: The resulting playbook must be actionable enough for an engineering and compliance team to implement immediately without a complete rewrite.

## Output format
Structure the playbook with clear Markdown headings for:
1. Executive Summary
2. Risk Classification Framework
3. Human-in-the-Loop & Override Protocols
4. Audit Trail & Logging Standards
5. Deployment Readiness Gates
6. Incident Response & Rollback Procedures

Customize with Prompt Like A Pro →