Automation often feels like the obvious answer to every repetitive task, but not everything should be automated. Some activities are better left manual, at least for now. To avoid wasting time and money on automations that do not move the needle, we use a simple three-question test before building anything.
1. Is This a Repetitive Task?
The first filter is repetition. Does this task happen over and over again in roughly the same way? Automation works best for activities such as invoice processing, status updates, report generation, or onboarding checklists. These tasks follow clear patterns and occur frequently.
If a task is highly variable, one-off, or constantly changing, automation usually creates more friction than value. Examples include unique client requests or strategic decisions. In these cases, it is better to keep the work manual. Save automation for the repeat offenders, the tasks that drain time and energy simply because they keep coming back.
2. Does It Cost More to Do Manually Than to Automate?
Next, consider the true cost of doing the task by hand. This includes hours spent, money involved, and the mental load it creates each month. If a process takes many hours and regularly introduces errors or delays, automation will often pay for itself quickly, even if it takes a few days to build.
On the other hand, if a task only takes half an hour once a quarter, the effort required to design, test, and maintain an automation will almost always outweigh the benefit. Focus on automating work where the manual cost is high and the return is obvious.
3. Will Automation Improve Quality or Just Speed?
The final and most important question is whether automation will make the outcome better, not just faster. Good automation reduces errors, such as automatically calculating invoices, or enforces consistency, such as standard onboarding steps. This kind of improvement frees people to focus on higher-value work.
If automation only saves time but does not improve quality, or even makes things worse, such as sending generic or tone-deaf automated messages, it is often better to keep the task manual or redesign the process first. Automation should raise the bar, not simply accelerate a flawed process.
When to Automate and When Not To
Use this three-question test as a quick sense check before starting any automation project.
Automate when the task is repetitive, expensive to do manually, and likely to improve quality or consistency.
Hold off on automation when the task is rare, highly variable, or when the main benefit is speed without better outcomes.
This simple filter helps teams avoid over-automating, under-delivering, and burning out on projects that do not truly matter.