A process needs redesigning rather than just automating when the steps themselves are inefficient, redundant, or exist for reasons nobody can explain. Automation is the right call when the process is sound but simply manual and repetitive.
The core distinction
A process needs redesigning, rather than simply automating, when the steps themselves are inefficient, redundant, or exist for reasons nobody currently working there can actually explain. Automation is the right call instead when the underlying process is fundamentally sound, but manual and repetitive. The logic works, it’s just slow and labour-intensive to carry out by hand.
A practical way to tell them apart
The clearest way to distinguish the two is to map the process first, before deciding which path to take. If the mapped-out process looks reasonable and logical but simply takes too much manual effort, automation is likely the right next step. If it looks convoluted, with steps that don’t obviously connect to a clear purpose, that’s a strong signal it needs redesigning before any automation is layered on top of it.
Can attempting automation reveal a redesign need?
Yes, and this happens more often than businesses expect. Trying to automate a process frequently exposes why it needed redesigning in the first place. Automation tools generally require clear, consistent logic to work properly, and a broken process resists being automated cleanly, forcing the underlying issues into the open.
Is it ever both, together?
Often, yes. The common and effective pattern is to redesign the process first to fix its underlying logic, and then automate the cleaned-up version. Attempting both simultaneously is possible but considerably riskier, since it becomes harder to tell whether a problem sits in the new process design or in how it’s been automated.
It’s worth adding that the answer isn’t always binary. Some processes benefit from a light redesign (removing one or two clearly unnecessary steps) followed by automating the streamlined version, rather than a full ground-up rebuild. Treating this as a spectrum, rather than a strict either-or choice, leads to a more proportionate response than defaulting to the more dramatic option every time.
Also worth knowing
- If nobody can explain why a step exists, it’s a redesign candidate, not an automation one.
- Redundant approvals or handoffs usually signal a redesign need.
- A sound process with repetitive manual steps is a good automation candidate.