How I decide what to automate

Before automating something, I try to describe what should happen when it goes wrong. If I can't do that yet, I probably need to understand the task better.

Stable, frequent chores are good candidates. Work that needs judgement can still benefit from a tool that gathers the information or prepares a draft. I can leave the decision to a person without making them repeat all the preparation.

I also need a way to tell whether the automation ran successfully. Otherwise I've given myself a new chore, checking on the tool.