Skip to content

1 · Spotting the tasks actually worth automating

Automation isn't about replacing yourself — it's about giving the boring, repeated parts of your day to a tool so you can spend your time on the things only you can do: talking to customers, making the product better, deciding what's next. But automating the wrong thing wastes money and creates messes you have to clean up. The first skill is knowing what to automate.

A task is a good automation candidate when it is:

  • Repetitive — you do it the same way over and over (sending the same reminder, copying numbers from one place to another, answering the same five questions).
  • Rule-based — there's a clear "if this, then that" with few surprises.
  • Boring and low-stakes — a small mistake is easy to catch and cheap to fix.
  • Eating real time — it actually adds up to hours, not two minutes a week.

Leave it to a human (for now) when the task is:

  • High-stakes — money decisions, legal stuff, anything where a mistake is expensive or hard to undo.
  • Full of judgment — an upset customer, a weird edge case, a sensitive message. People can tell when they've been handed to a machine on something that mattered.
  • Rare or always different — automating something you do twice a year usually costs more than it saves.
  • About relationships — the personal touch is the value; don't automate it away.

This is exactly the discipline the U.S. Small Business Administration points small owners toward: master your day-to-day operations deliberately, choosing where technology helps you run leaner (SBA, n.d.). SCORE — a nonprofit that mentors small businesses — frames automation the same way: automate repetitive tasks so you can focus on the work that grows the business, not so you can take humans out of the parts that need them (SCORE, n.d.).

Check yourself. Name two traits that make a task a good automation candidate, and one kind of task you should keep a human on even if you could automate it.

Sources