Skip to content

Note

When automation is the wrong answer

Automation is useful when a careful task keeps returning. It is less useful when someone is trying to avoid making a decision.

Most requests for automation are sound. A person is doing something exact, by hand, many times a week, and a machine could do it more patiently. Reading a form. Moving a record. Checking one rule. That work should be automated, and the person should get their afternoon back.

Some requests are different. The task is not repetitive so much as unpleasant. It involves a judgement nobody wants to own, or a conversation nobody wants to have. Automating it does not remove the difficulty; it hides it, and adds a layer of software that now has to be maintained around the thing everyone was already avoiding.

We try to tell the two apart before we build anything. If the honest description of the problem is “this is careful, repetitive, and clear,” automation is a gift. If it is “this is a decision we keep deferring,” the useful work is helping someone make the decision, not building a system that pretends it has been made.

A good automation makes a real task faster and leaves the responsibility exactly where it was.

Start a project

Tell us about the system you need to get right.

A senior engineer replies within two working days with a straight view of scope, risk and the fastest sensible path — no sales layer in between.