When Tunde handed in his resignation, his manager was baffled. He was good at his job, well liked, and paid fairly. In the exit conversation, the real reason came out: "I didn't train as an accountant to spend three days every month copying numbers between a bank statement and a spreadsheet. It's soul-destroying, and I can't do it anymore."
His manager did some quick maths. Three days a month, every month, from a skilled member of staff — nearly forty working days a year spent on a task that produced no insight, just moved data from one place to another. And now it had cost them a good employee too.
Here's the part that stings: that task didn't need a person at all. It was the perfect candidate for automation, hiding in plain sight, quietly draining time and morale for years. Most organizations have several of these. The trick is learning to see them.
Every business runs on processes — and processes leak
Approvals, data entry, reconciliations, reporting, onboarding: these are the machinery of a company. When they run smoothly, nobody notices. When they run by hand, they leak — time, accuracy, and often the goodwill of the people stuck doing them.
The good news is that the automation which plugs these leaks usually pays for itself faster than almost any other investment you can make. You're not betting on a new market or a new product; you're removing a cost you're already paying, every single month.
The only real question is which processes to start with. Get that right and the returns are obvious and fast. Get it wrong — automating something rare, or something that genuinely needs human judgement — and you spend a lot of effort for little payoff.
The three traits of a process worth automating
The best candidates almost always share three traits. Tunde's reconciliation task had all of them.
- High volume. It happens hundreds or thousands of times a month, or eats whole days on a fixed cycle. Volume is what turns a small saving per task into a large saving per year.
- Rules-based. A knowledgeable person could write down how each step is decided. If the logic can be described — "match this to that; if the amounts agree within a naira, clear it" — software can usually handle it.
- Costly when wrong. Mistakes create rework, delays, or unhappy customers. A missed duplicate payment or a late report has a real price, and removing that risk is part of the return.
Hold your manual processes up against those three. The ones that tick all the boxes are where automation earns its keep first. The ones that tick none are best left to people — no matter how tempting they look on a slide.
Count the cost you can't see
The reason these opportunities go unnoticed is that manual work hides its price. It's spread thin across people's days, buried inside "that's just how we do it."
So make it visible. For a candidate process, add up honestly:
- The hours — how long it takes, times how often it happens.
- The errors — how often something has to be corrected, and what each correction costs downstream.
- The delay — what it costs to have the output late, whether that's a report leadership needs or a customer waiting.
- The ceiling — the growth you can't take on because this process would need more people to scale.
That total — hours, errors, delay, and the growth you're leaving on the table — is what automation gives back. In Tunde's case, add the cost of replacing him, and the business case wrote itself.
Start small, then let it compound
The instinct, once you see the opportunity, is to automate everything at once. Resist it. The organizations that get this right pick one well-scoped, high-return process, ship it properly, and prove the saving before moving on.
There's a reason beyond risk management. When people watch one genuine chore disappear — when the finance team gets three days a month back and stops dreading the close — something shifts. Scepticism turns into enthusiasm, and colleagues start bringing you the next candidate themselves. The programme stops being something done to the team and becomes something they pull toward.
That momentum is worth more than any single automation. It's how a first win becomes a second, and a second becomes a habit.
The version of this story that ends well
Tunde's company automated the reconciliation after he left. The task that drove him out now runs quietly in the background; a person reviews only the handful of exceptions that genuinely need a human eye. The irony isn't lost on anyone: had they done it a year earlier, they'd probably still have him.
You almost certainly have a task like his — repetitive, rules-based, quietly expensive, and grinding down whoever's assigned to it. Find it, count what it really costs, automate it properly, and measure what comes back.
Curious where automation would pay off first in your business? We're happy to help you find it. Explore our automation services or book a consultation.
