Mission
A mission is a recurring or one-time job assigned to an AI Employee — a unit of work with a goal, inputs, and a definition of done.
Where 'task' suggests something small, 'mission' carries the weight of an objective with a clear outcome. A mission has: a goal, a trigger (manual, scheduled, or event), inputs, tools the employee may use, an approval gate, and a definition of done. Missions are the right abstraction for assigning work because they survive the task/agent gap: a single mission can spawn many runs, call many agents, and span days. The employee owns the mission; the platform tracks its history.
Example
Mission: 'Send weekly digest to top 50 customers'. Trigger: every Friday 9am. Inputs: customer list, last week's product changes. Tools: email, analytics. Approval gate: founder reviews drafts before send. Done: digest sent, opens logged.
How OpenLabor uses it
Missions are first-class in OpenLabor. You define them once; the employee runs them on schedule with a clear audit trail.
How is a mission different from a workflow?
A workflow is the recipe. A mission is the standing order to execute that recipe — with a goal, an owner, and history.
Can missions trigger other missions?
Yes. A 'qualify lead' mission can trigger a 'book meeting' mission. Chaining is normal; runaway chains are why you set bounds.
Related: run, ai-employee, approval-gate, agentic-workflow.
AI Labor Glossary