When an assignee finishes a task, it moves to In Review and lands on your queue (the Assigned by me or Team scope on My Tasks; All tasks if you may view every task). Reviewing closes the loop — and feeds KPI.
Review the output
Open the task and read the Output tab — the assignee's answer (plus any output files). The original ask is on Overview.
Rate it → Done
In the Review the output panel, pick a grade and press Submit rating:
- Exceeds Expectations
- Meets Expectations (the default)
- Needs Improvement — a written reason is required for this one.
That closes the task as Done and records the grade.
…or send it back
Not good enough yet? Write a note in What needs to change? and press Request changes. The task returns to In Progress for the assignee, with your note — no penalty, just a loop.
The 48-hour window
If nobody rates a task within 48 hours of it being submitted, it auto-approves as Meets so work never gets stuck on an absent manager. This is a safety net, not a verdict. The Orchestrator warns you (about 12 hours before) when a review is about to auto-pass, so you can weigh in first.
Changing a grade later
Once the task is Done the panel is headed Grade and the button reads Override grade. You can change the grade of an already-closed task (a downgrade to Needs Improvement still needs a reason). It is audited.
The assignee's response
The person who was graded sees a Your response to this grade panel and can press Acknowledge grade, optionally with a comment. This does not change the grade or reopen the task — it records their side in the task's permanent history.
Sub-tasks roll up
If the task had dependent sub-tasks, completing them rolls their grades up into the parent automatically — you do not rate the parent separately. See Sub-tasks & delegation: how work rolls up the tree.
Related: Task KPI and grading: how it works and the fairness rules · The Task Orchestrator: keep work moving while you are away.