What should I automate first?What to automate first.
5 minute read
Updated
Most people pick wrong here, and they pick wrong in a predictable way: they choose the task they personally hate most. That is rarely the task that pays back first.
The thing worth automating first is boring, frequent, and follows rules you could write down. If you get that choice right the first project pays for itself and everyone in the business becomes willing to try the second one. Get it wrong and you spend months on something clever that nobody uses.
The three-question test
Take any task and ask these in order. A job that passes all three is a strong first candidate.
- How often does it happen? Daily beats weekly, weekly beats monthly. Something that happens twice a year will never repay the work of automating it, however annoying it is.
- Could a new hire do it from written instructions? If yes, a system can do it. If the honest answer is "you have to have been here a while", it needs judgment and it is not first.
- What happens if it is done slightly wrong? Low stakes means you can launch, watch, and improve. High stakes means heavier testing, which is fine later but slow for a first project.
The jobs that usually pay back first
Across almost every small business, the same few keep appearing. If one of these is true for you, start there.
- Answering the same handful of questions. Opening hours, pricing basics, whether you cover their area, what happens next.
- The phone going unanswered while you are doing the work someone already paid for.
- Retyping the same information from one system into another.
- The weekly report someone assembles by hand from three places.
- The follow-up that should happen and does not, because the person who should send it is busy.
What not to automate first
These are the ones that look appealing and go badly as an opening move.
- Anything where nobody agrees on the rules. If two people in your business would handle it differently, the disagreement is the actual problem and automating it just makes it faster and wrong.
- The rare and complicated thing. High effort, almost no payback, and it is usually rare precisely because it needs a person.
- Anything that only your most senior person can do. That is judgment work, and it is the last thing to touch, not the first.
- A task you are about to change anyway. Automate the process you are keeping, not the one you are replacing next quarter.
Work out the payback before you buy
You do not need a spreadsheet. Take how many times the task happens a week, multiply by roughly how long it takes, and you have hours a week. Then ask what those hours are currently costing you, either in wages or in work you are not doing because you are doing this instead.
That number is what you are buying back. If it is small, this is not the right first project, and a vendor worth trusting will tell you so rather than quote you anyway.
A reasonable first move
Pick the job that happens most often and needs the least judgment. It will feel unambitious. That is the point: it goes live quickly, it works, and it earns you the internal permission to do something bigger.
For most businesses that is either the questions arriving faster than anyone can answer them, or the calls nobody can pick up. Both are frequent, both follow rules, and both cost you customers quietly rather than loudly.
On a call we will do this exercise with you against your actual week, and if nothing clears the bar we will say so.
Still not sure which applies to you?
Fifteen minutes on a call is usually enough to settle it, and we will tell you if the answer is something we do not sell.
Book a callFAQ
Common follow-up questions.
Ready to put AI to work?
A free 15-minute AI consultation. We learn your business, you learn whether this is a fit, and your quote is confirmed on the call.