How OtterDesk works

OtterDesk helps you supervise AI co-workers without treating them as a black box. Each co-worker is a role you add to your team, such as monitoring, research, operations, or another focused responsibility.

Co-workers

A co-worker has a mission, configuration, activity history, and an optional schedule. You can give it a recognizable name, test it, pause it, review it, or archive it when you no longer need it.

Missions and work

The mission explains what the co-worker is responsible for. Work happens in individual runs. A run may be started by you with Test now or by a schedule.

Evidence and decisions

OtterDesk keeps activity and evidence with the co-worker's work. If a co-worker needs a human choice, it shows a Decision needed card. It waits for your response instead of acting past that boundary.

Trust

Trust is practical, not assumed. Before relying on a co-worker, inspect:

• Its mission and expected input and output.
• Its connected data sources and stated limitations.
• Its recent activity and evidence.
• Its approval requirements.
• Its resource requirements and current health.

Use Review work for the day-to-day review process and Privacy and safety for the data and control boundaries.