Complexity is the amount that the steps of a task change from one instance to another – irrespective of number of steps. An example is the processing of invoices from a single vendor which are unlikely to change from invoice to invoice. On the other side of that coin, the processing of all the documents defining a relationship with a vendor (which includes invoices as well as financials, POs, LOA, and correspondence) is complicated and may require previously undefined steps often. Preparation of tax schedules that may reference checks, websites, and statements from any number of providers is another example of a task with steps that can change from instance to instance.
Current level of digital processing represents the portion of the steps that are processed by the computer. Anything that requires paper obviously possesses a low level of digital processing. But reading scanned handwritten documents is another process that requires human intervention even though it is done on a computer.
It is not surprising that some processes are not candidates for RPA. Anything in Quadrant II where the process is manual, changes with each instance, and requires expertise and human intuition will not automate – even when done on a computer. A clear example is custom software development where an individual must think through the solution and test that it does what was intended.
Quadrant III includes processes that may go either way. Clearly the manual steps will need to be processed by a human, but the static nature of the activity suggests that some automation may be possible. An example of a Quadrant III task that can’t be automated by RPA is picking grapes (a duck walks into a bar…). But a process of recording retail shrinkage (theft and spoilage) which involves physical measuring of inventory followed by computer processing, distribution, and recording could benefit from some RPA.
Quadrant I offers additional opportunities for semi-automation. Tasks in this area are fully digital but require lots of manual intervention. Both of my examples of complication above, preparing tax schedules and defining a relationship that includes many types of digital documents are examples that can benefit from semi-automation.
Quadrant IV is where RPA really shines. Regardless how many steps are included in each task or how many tasks are included in each process, each component can be easily defined and does not change from instance to instance. Returning to my example above, it does not matter how many different systems are involved in the onboarding of a new employees, if they are all accessible from one computer and are the same for every new hire, an RPA robot can be used to complete the entire task. Do you have processes in your organization that meet the criteria of Quadrant IV?