A company should automate a workflow when the steps are repeatable and agreed, the volume is high enough for time savings to matter, the inputs are available in digital form, and the cost of an occasional error is manageable or can be caught by review. If the process is still changing or nobody owns it, fix that first.
Five questions to ask
Four or five 'yes' answers suggest a good candidate. Two or fewer suggest the process needs work before automation.
- Do the steps happen the same way most of the time?
- Is there enough volume that saving a few minutes per case adds up?
- Is the information needed available digitally, or could it be?
- Is there a clear owner for the process and each step?
- If the automation gets something wrong, will someone notice and can it be corrected?
Signs a workflow is not ready
- People disagree about how it should work
- It happens a few times a month
- Every case is an exception
- Errors would have serious legal, financial or safety consequences and would be hard to detect
How to choose what to automate first
Score candidate workflows by volume, time per case, cost of delay and risk. Start with one that is high in volume and delay cost but low in risk. Measure a baseline before changing anything, so you can show whether the automation helped.
Limitations and considerations
- Volume and risk estimates are only as good as the data behind them; measure where possible.
- Automation creates maintenance work when connected systems change.