A clear definition

Automation

Automation is the use of a system to perform defined work with reduced manual intervention.

What it means in practice

It can use fixed rules, integrations or AI, depending on the task. The important questions are whether inputs are reliable, the desired output is clear and exceptions have a responsible owner. Count checking, correction and maintenance when assessing value. A faster process can still be a poor process. Match the control to the consequence of error and preserve a recovery route when the system fails. Assess the complete operating effort after introduction, including review, exception handling and maintenance. Time released does not automatically become payroll saved. A workflow should have a responsible owner and a manual recovery route. For consequential actions, distinguish preparing an output from authorising it. The value of automation depends on reliable work and useful capacity, not merely on removing a visible human step.

A fictional worked example

A system can prepare a weekly report from approved records while a person reviews unusual values before sharing it.

A useful question

What happens when the normal input is missing, contradictory or repeated?

Read the practical guide