What is one task you would trust an AI agent to do without checking?

I think the interesting question with AI agents isn't what they can do. It's what people are actually comfortable letting them do on their own. For example: Sending routine emails Updating a spreadsheet Scheduling meetings Monitoring something Following up with customers Sorting incoming requests What's one task you would genuinely let an AI agent handle without reviewing every step? And what makes you trust it enough to do that? I'm curious where people draw the line.