Insights

Which Manual Processes Are Worth Automating

Not every repetitive task should be automated. A simple way to decide which ones pay for themselves and which ones will not.

Automation gets pitched as though the only question is whether a task is repetitive. It is not. Plenty of repetitive work should stay manual, and automating it makes things worse in ways that show up later.

The cost nobody quotes

An automated process has a build cost, which everyone counts, and a maintenance cost, which almost nobody does. It has to keep working when the form changes, when the file arrives in a different format, when the exception nobody anticipated turns up. Automation converts a small recurring human cost into a smaller recurring human cost plus an occasional large one.

That trade is often worth making. It is not always.

Where automation pays back

  • The rules are stable. If the logic has not changed in a year, it probably will not change next year either.
  • Volume is high enough that mistakes are certain. A human doing something forty times a day will get it wrong sometimes. That is not a discipline problem.
  • The delay is the cost. Work that waits for someone to be at their desk — approvals, handoffs, notifications — often costs more in latency than in labour.
  • It has to be auditable. When someone will eventually ask who did what and when, a system answers that and a person does not.

Where it usually does not

  • The process is still changing. Automating an unsettled process fixes it in place before anyone knows whether it is right.
  • Every case is an exception. If the rules are really a series of judgement calls, automation produces a system that handles the easy 60% and hides the hard 40%.
  • It runs a few times a month. Ten minutes of manual work twice a month does not justify a system somebody has to maintain for years.
  • The manual step is where the thinking happens. Some tedious-looking steps are where someone notices things are wrong. Automate that and you lose the check without noticing.

A useful order

Before automating, try to remove the step. A surprising number of recurring tasks exist because of something upstream that could be fixed instead — a form collecting a field nobody needs, a report nobody reads, an approval nobody has ever declined.

Automating a process you should have deleted is the most expensive outcome available, and it is common, because the automation makes the waste invisible rather than absent.

Recognise any of this?

Start with a conversation. We will help you scope the problem before anything gets built.

Start a Project