The Human-AI Delegation Map gives a team a shared answer to a question that tool policies often leave vague: who owns which part of the work? It turns broad AI guidance into task-level boundaries people can use every day.
Map responsibility before adding agents
Place recurring tasks into four columns: work that stays human-owned, work AI may draft, work that requires named human review, and work the team will not delegate. Discuss why each boundary exists and what evidence would justify changing it.
The result is a practical human-in-the-loop agreement. It helps leaders delegate to agents without hiding accountability or forcing every employee to invent their own rules.
How to facilitate
- 1Map tasks, not job titles. Roles contain work with different levels of risk and judgment.
- 2Name the reviewer. “Human review” only works when ownership is clear.
- 3Consider reversibility. High-impact or hard-to-reverse decisions need stronger human control.
- 4Review monthly. Update boundaries as tools, evidence, and regulations change.
When to use this template
Use this map before deploying AI agents, redesigning a role, automating customer communication, changing a product workflow, or creating an internal AI policy.
