The fastest way to be disappointed by automation is to start with the wrong job. Pick something fiddly, rare or low-value and you'll spend money for a result nobody notices. Pick the right one and it pays for itself in weeks.
So here's a simple way to choose — no spreadsheets of scoring required.
Look for jobs that are frequent, repetitive and rule-shaped
The best candidates share three traits:
- Frequent — it happens daily or weekly, not once a quarter. Frequency is where the savings add up.
- Repetitive — it's basically the same steps every time.
- Rule-shaped — you could explain how to do it in a few clear instructions.
A job with all three — raising invoices, re-keying documents, building the weekly report — is automation gold. A job that's rare, or different every time, or needs real judgement, usually isn't worth automating first (and sometimes not at all).
Start with the one that hurts most
Among the frequent, repetitive jobs, pick the one your team dreads — the Friday-afternoon report, the end-of-month scramble, the inbox that never empties. Two reasons: the pain makes the payback obvious, and a quick win builds the confidence to do more.
We always start a build with the job that hurts most rather than the one that's technically tidiest. The first "that's just done now?" moment matters more than perfection.
Do the rough maths
You don't need a formal business case, just an honest estimate:
- How many hours a week does this job eat?
- What does that time cost — and what does not having it cost (late invoices, missed deadlines, mistakes)?
If a job swallows a day a week, that's a fifth of someone's time. Automating it for a fraction of that cost, paying back within months, is an easy yes. If the maths is marginal, leave it — there's almost always a better first job. (We dig into this in how much automation costs.)
Avoid the classic mistakes
- Don't automate a broken process. If a job is a mess done by hand, automating it just makes the mess faster. Tidy the process first, or fix it as part of the build.
- Don't start with the rare, complicated one. It's tempting because it's annoying — but rare means low payback and high effort.
- Don't try to do everything at once. One job, proven, then the next.
Let the audit do the choosing
Honestly, the easiest way to choose well is to talk it through with someone who's seen it before. That's the whole point of our free 20-minute audit: we look at where your week actually goes, point at the jobs worth automating first, and tell you plainly which ones aren't worth it. You can see the kinds of work we take on across our services.
Choose the right first job and automation stops being an expense and starts being the best hire you've made all year.