Automation is easy to start and hard to stop. The temptation is to automate everything that can be automated — which is how you end up maintaining a pile of brittle scripts that each save five minutes and collectively cost a week.
Where to start, where to stop
Automate the work that is frequent, stable, and rule-based — the same boring thing, done often, the same way. Leave alone the work that's rare, that changes constantly, or that needs judgment; automating those costs more to build and maintain than they'll ever save, and tends to break in ways no one notices until it matters.
Good automation candidates are:
- Frequent enough that the time saved is actually real.
- Stable enough that you're not rewriting the automation every month.
- Rule-based enough that a machine can do it without judgment calls.
The question isn't what can be automated. It's what's worth automating — and worth maintaining once you have.
Every automation is a small system that someone now owns forever. Build the few that clearly earn their keep, skip the many that don't, and you'll spend your time running the business instead of maintaining the machinery you built to run it.