Scaling up Robotic Process Automation (RPA) adoption - part 2
Which process selection approach is better?
Why isn’t RPA spreading like wildfire throughout every organization? It could be the way processes are selected for automation. My previous post identified two approaches to process selection. Are you looking from the bottom-up, at individual IT tasks? Or are you looking from the top-down, at business processes?
Bottom Up: Typically, you identify a task and automate it. These tasks are performed by individuals and likely will be part of a larger process. A task is smaller units of repetitive, routine steps performed to support the larger function. It is smaller than a subprocess. Example: A routine expense entry into an ERP system by an individual. Tasks that can be crowdsourced means they can easily be identified by employees. Tasks can be identified by socializing the need for automation and by providing a mechanism to register ideas (tasks) for automation. Since individuals perform tasks, typically they are not the same tasks, so the impact of the automation is limited to individual effort savings. Individuals can have limited, routine tasks that can be automated.
Top-line: Typically, you identify an entire business process, optimize it, identify sub-processes/tasks for automation, calculate ROI and automate it, if ROI is good. For example, the expense approval business process. It can be a subprocess, such as approving and processing expenses (payments). The impact of automation is not only limited to individuals but also functions on the department level as backend processing. Processes make a significant difference in scale, speed, and productivity and accelerate achieving the organization's goals.
I’ve outlined the advantages and disadvantages of the bottom-up and top-down approaches in this chart.
With either of the approaches, you cannot unearth all of the automation opportunities. You need both approaches, bottom-up (task-oriented) and top-down (business process-oriented). A well-coordinated RPA program allows for a 30% to 40% productivity improvement by a reduction in execution time, an increase in speed, and increased accuracy.
Let's look at how these approaches are useful in the next post.