A four-question test for whether a job can be automated

People usually come to me with either “can AI do everything?” or “I don’t think any of this applies to me.” Both are wrong, and there’s a quick way to find out which jobs are actually worth it.

Take one job in your week and ask these four things.

1. Does it repeat on a rhythm?

Every week, every month, every time an enquiry comes in. If it’s genuinely one-off, automation is the wrong tool — you’d spend more time building it than doing it.

The catch: people badly underestimate what repeats. “I don’t really have repeating tasks” is almost never true. Write down what you did last Friday and the Friday before. Things will start looking familiar.

2. Could you write down the steps?

Not perfectly. Roughly. “I open the reports, I copy the numbers into the sheet, I compare to last month, I write a paragraph about what changed.”

If you can say it out loud in order, it can be built. If you can’t — if it’s genuinely judgement, taste, or knowing your customer — that part stays human. Most jobs are a mix, and the good move is automating the boring three-quarters and keeping the judgement for yourself.

3. Is the information somewhere a computer can reach?

This is the one that quietly kills projects. If the numbers live in a system with a login, an email, a spreadsheet, a website — fine. If they live in a filing cabinet, or in your head, or in a photo of a receipt in your phone, then something has to change first.

Not a dealbreaker. Just work that needs doing before the interesting part.

4. What happens if it gets it slightly wrong?

Be honest here. If a draft goes out with a clumsy sentence, you shrug. If an invoice goes to the wrong client, that’s a phone call you don’t want to make.

This doesn’t decide whether to automate — it decides how much human checking sits in the middle. Low stakes, let it run. High stakes, it prepares everything and waits for you to press yes. Both are automation. Only one should run unattended.

Scoring it

Four yeses? Build it. That’s a good first project and you’ll feel the difference within a month.

Three yeses with a wobble on question three? Still worth it, there’s just some tidying first.

Two or fewer? Leave it. Pick a different job. There’s almost always a better candidate sitting right next to it.

The one thing people get backwards

They pick the biggest job. The month-end monster, the thing that eats two days.

Don’t. Pick the most annoying small one. It’ll get built quickly, you’ll see it working, and you’ll actually trust the next one. Confidence matters more than hours saved on the first project — because the first project is really about finding out whether this works for you at all.

If you want a second opinion on which job to pick, that’s genuinely what my free consultation is for. Bring me the list and we’ll sort it together.