AI agent demos keep getting more ambitious.
But I suspect people will learn to trust agents through boring, low-risk tasks—not by letting them “run an entire business.”
Things like:
• rescheduling a meeting
• reordering an inexpensive item
• comparing travel options
• sending a follow-up based on meeting notes
• monitoring a price and notifying you when it changes
For me, the key question isn’t how complex the task is.
It’s whether the action is reversible and whether a mistake would be expensive.
What’s one task you’d let an AI complete without checking first?
And where would you still want it to stop and ask?
[link] [comments]