5 questions to find good RPA candidates
We have a lot of interest at my company to implement Robot Process Automation (RPA). One of the things I've started doing is meeting with different HR leaders to talk about different processes that could be good candidates for RPA. In some of the initial discussions, I hadn't established a set of traits or rules for evalutating whether these were good use cases or not.
Is the current process standardized?
Don't build the plane while you're flying it. Also don't forget that digitizing a crappy process often returns a crappy digital process.
- Is the process well-understood?
- Is the current process already optimized?
- Are there clear rules and decision points in each step of the process?
- How often or when does the process change?
Is the data well-structured?
If you can't read or place the same data in the same place every time the process runs, how can you automate it? (Hopefully this will be less of an issue one day with AI.)
Does the process run frequently?
There's a whole lot more value in automating a 5 minute process running every hour than a 90 minute process running once a month.
Is there significant manual effort?
This question tells us the complexity to build the RPA process and the potential value of automating the manual process.
- How much manual effort is spent on the current process?
- How many systems are involved?
Are there many exceptions?
How many exceptions or special scenarios are there? More exceptions = less automation effectiveness and/or more development time to handle those scenarios.
- How are exceptions handled?
- Who will manage when the automation fails?