Automation pays off fastest on the small, repetitive tasks nobody enjoys — not the ambitious system that takes a year. Here is where to start, and how to tell a good candidate from a bad one.
Automation has a reputation for being ambitious and expensive. The version that actually pays for itself is neither. It is the quiet removal of small, repetitive tasks that eat an hour here and an hour there and never get faster no matter how experienced the person doing them is. Start there, not with the grand system.
What makes a task worth automating
A good candidate has three qualities: it happens often, it follows the same steps every time, and a person currently does it by hand. A task that happens twice a year is not worth the setup. A task that needs judgement every time cannot be automated without making it worse. The sweet spot is the boring, frequent, rule-following work — and every business has more of it than it thinks.
Where to look first
- Following up on quotes and unpaid invoices. A polite reminder sent on a schedule recovers money that otherwise leaks away because nobody had time to chase it.
- Re-typing the same data into two systems. Anywhere a person copies details from an email into a spreadsheet, or from an order into an accounts package, is a place a machine should be doing the copying.
- Appointment and deadline reminders. No-shows and missed dates cost real money, and a reminder that goes out automatically the day before costs nothing after it is set up.
- Assembling the same report every week or month. If someone spends the first hour of Monday pulling numbers into the same layout, that hour is recoverable.
- Onboarding a new customer or staff member. The welcome email, the account setup, the checklist — the same sequence every time is exactly what automation is for.
The honest test
If you can describe a task as "every time X happens, someone does Y", it is probably automatable. If describing it needs the words "it depends", it probably is not — at least not yet.
Settle the process before you automate it
Automation makes a process faster, including a broken one. If two people do the same task differently, automating it forces the disagreement into the open — usually at the worst moment, mid-build. Agree the single right way to do the task first, in writing, then automate that. It takes an afternoon and saves a fortnight.
Then expand where it connects
Once the first task runs reliably, the next one in the same area is cheaper to automate, because the connections and access are already in place. Automation compounds: the second win in a domain costs less than the first. That is the argument for starting small and specific rather than trying to boil the whole business in one project.