To check if a process or a workstream is a good candidate for being automated, see if it meets all three of these criteria:
- The process must be a well-oiled machine. The requirements and outcomes are well established. Is the process stable enough to be automated?
- The process doesn’t need someone to engage with it each time. It doesn’t need manual intervention, oversight, excessive customization, or finesse. It runs in the backdrop; it’s boring and doesn’t require ‘higher-order’ thinking. Are there decision points within the process that require human intervention?
- The process is time-consuming. By automating it, will you save at least 4x what you’ll invest in automating it?
If the manual process is broken or doesn’t exist, then automating it before it’s a “well-oiled machine” may lead to mistakes and unnecessary rework. Establish success with the manual workflow before attempting to automate it.
Idea for Impact: Picking which processes to automate isn’t easy; yet, the closer you observe the workflow deeply, the sooner you can understand both the happy path to automation and the exceptions.
Alfred F says
NEVER automate something that can be eliminated (see 80/20 rule)