Which admin tasks are actually worth automating
The usual advice is to automate everything repetitive, which is wrong and expensive. Here is a way to work out which tasks actually pay back, and which to leave alone.
The standard advice on automation is to find repetitive tasks and eliminate them. It sounds obviously correct and it leads people to spend real money automating things that were never costing them much.
Every automation has a build cost and a permanent maintenance cost. Some tasks clear that bar comfortably. Others do not, and the honest answer for those is to keep doing them by hand.
The arithmetic
Work out the annual cost of the task as it stands: how often it happens, multiplied by how long it takes, multiplied by the hourly cost of whoever does it. Then add the cost of the mistakes it causes, which is often the larger number.
Two years is a reasonable payback for something that keeps working. Two years for a task that will change in six months is not.
Frequency matters more than duration
People instinctively want to automate the task that feels most painful, which is usually a long one. That instinct is often wrong.
A tedious quarterly job taking three hours costs twelve hours a year. A two minute task done fifteen times a day costs over a hundred hours. The second one barely registers as a burden because each instance is trivial, and it is by far the more expensive problem.
Frequent small tasks also carry a hidden cost that does not show up in the arithmetic: each one interrupts something else. The two minutes is rarely just two minutes.
What usually is worth automating
- Moving data between systems. High frequency, entirely rule based, and error prone precisely because it is boring. Almost always the best place to start. There is more in when to connect your existing tools.
- Reminders and follow-ups. They must happen on time, they are trivial to define, and human memory is the least reliable way to deliver them. Appointment reminders are the clearest case, covered in reducing no-shows.
- Recurring reports. If someone assembles the same figures every week, that is a report the system should produce.
- Document generation. Quotes, invoices and certificates built from data you already hold. Fast to automate and it removes a common source of embarrassing errors.
- Status updates to customers. Repetitive, expected, and a direct saving on inbound queries.
What usually is not
- Anything requiring judgement. If the answer depends on context a person weighs up, automating it produces confident wrong answers rather than saved time.
- Genuinely rare tasks. Something done twice a year does not justify a permanent piece of software, however irritating it is.
- Processes that change constantly. If the rules shift every few months, the automation becomes a maintenance burden that outweighs the saving.
- Work that is really a conversation. Some emails look repetitive but are actually relationship management. Automating them saves minutes and costs goodwill.
- Anything you are about to replace. Automating a workflow inside a system you plan to move off is money spent twice.
The trap worth avoiding
The most common expensive mistake is automating a process that should not exist.
Businesses accumulate steps that made sense once. A form filled in because a manager wanted it years ago. A spreadsheet updated for a report nobody reads. A double entry that exists because two departments did not trust each other in 2019.
Automating these makes them permanent. They stop being visible, so nobody ever questions them again. Before automating anything, ask what would happen if the step simply stopped. Occasionally the honest answer is nothing at all, and you have just saved the entire build cost.
The best automation is deleting the step. The second best is automating it. Doing it faster by hand is a distant third.
How to choose the first one
Do not start with the most impressive candidate. Start with one that is frequent, rule based, and contained within a single system if possible.
- Log every repetitive task for a week, with how long each takes. Include the tasks nobody thinks of as tasks.
- Rank by annual hours, not by how annoying they feel.
- For the top three, ask whether the step needs to exist at all.
- Of what remains, pick the one with the clearest rules. Clear rules mean a cheap build and a reliable result.
- Measure the hours recovered, then use that figure to decide whether a second automation is worth funding.
That last step is the one people skip, and it is what turns automation from an act of faith into a decision you can defend.
Frequently asked questions
Multiply how often it happens by how long it takes by the hourly cost of the person doing it, then add the cost of errors it causes. Compare that annual figure against the build cost plus ongoing maintenance. A payback period under two years is generally worth doing, provided the process is stable.
After, in most cases. Automating a workflow inside a system you are about to replace means paying for the same work twice. The exception is a standalone automation that will survive the change, such as one that talks to your accounting software rather than the system being replaced.
Automating a process that should have been deleted. Businesses accumulate steps that made sense once and persist out of habit. Automating them makes them permanent and invisible. Always ask what would break if the step simply stopped before building anything.