Automation
Automate when a process is repetitive, rule-based, high-volume and stable, and when the steps have already been simplified. Automating a process that is still changing shape locks in the version you were about to improve.
The task people complain about most is often infrequent and full of judgement: a poor candidate. The task nobody mentions because it is merely tedious, done forty times a week identically, is usually the better one.
The most valuable output of mapping a process is normally the discovery that several steps exist for reasons that no longer apply. Removing a step returns more than accelerating it, and it is cheaper.
Most processes have a stable majority and a long tail of special cases. Handling the majority automatically and routing exceptions to a person with full context captures most of the benefit; chasing total coverage is where cost and fragility escalate.
Do not automate what should stay accountable.
Pricing a difficult deal, resolving a complaint, deciding an unusual case: these belong with people. Automating them tends to produce confident wrong answers and removes the judgement that made the outcome defensible.
Low frequency, unstable steps, heavy judgement, or a process currently being redesigned. Automating any of those costs more to build and maintain than the time it returns.
Yes, and expect the map to change the plan. Mapping usually reveals steps that can be removed outright, which returns more value than automating them would have.
Automate when a process is repetitive, rule-based, high-volume and stable, and when the steps have already been simplified. Automating a process that is still changing shape locks in the version you were about to improve.
You have read what we think. If you want to know what it means for your case specifically, describe it and we will tell you which parts of when should you automate a process? actually apply — and which do not.
