Question 15
Domain 2A team boasts that their Cowork task ran autonomously for a week with no human interaction. What's the structural critique?
Correct answer: A
Explanation
This is the “set-and-forget antipattern” because Cowork is “collaborative-by-design,” so a task left alone for a week violates the intended operating model. The critique is that “long-running agents drift silently without checkpoint cadence,” which is why the right pattern is “set-and-steer with daily checkpoints.”
Why each option is right or wrong
A. Set-and-forget antipattern. Cowork is collaborative-by-design; long-running agents drift silently without checkpoint cadence. The right shape is set-and-steer with daily checkpoints.
The structural flaw is the absence of the required checkpoint cadence: a week-long autonomous run means the agent was left to drift without any steering points, which is exactly the failure mode described as set-and-forget. In Cowork’s operating model, collaboration is not optional; the task should be re-anchored at regular intervals, typically daily checkpoints, rather than allowed to run unattended for 7 days.
B. Cowork doesn't support week-long tasks.
This claims a hard product limitation, but the issue is governance and collaboration structure.
C. The agent should self-monitor.
Self-monitoring helps, but it does not replace human checkpoints for alignment and oversight.
D. Tasks should always be under 1 hour.
This invents an absolute time threshold; task duration depends on risk and checkpoint cadence.