We have heard all the good and bad about RPA. The good, RPA can reduce the existing system cost of ownership up to 40%. The bad, up to 50% of RPA projects fail.
Given there are success stories that are amazing in RPA like the one Chet Chambers from UiPath discusses, where he automated over 500 processes before he left HP, it’s no secret that RPA should be in every company’s future. So let’s talk about how to achieve that success and more specifically how you can take the proper steps to achieve the efficiency that RPA provides.
Keep in mind that RPA is not digital transformation, it is taking existing processes and the underlying applications and then applying automation where humans would have typically been interacting with the system. This usually includes doing comparisons, excel extracts and other types of interactions that are routine. With this said, here are the beginning steps:
1.Understanding that RPA is not meant for highly complex operations (for that see Machine Learning), we can now understand how to map out what processes we should be taking on and a way to paint a broader picture. Most companies are able to automate a few RPA processes but stall out due to lack of vision and organization.
When we do RPA programs we create a project plan, just like any other software rollout, along with artifacts. Common artifacts will be an end state definition, Process Definition Document, Process Repository and StakeHolder objectives document. Having these filled out and ready to go day one will create a path of success for the organization. If you don’t have these documents you have no way to quantify success.
2.Once you have automated your initial processes it is important to find and cultivate your second champion. This is the team or individual that you will move to once you have your low hanging fruit already accounted for and created. Usually you have to scour the company and educate people about RPA in order to build a funnel of automation opportunities.
This is a proactive activity that should be happening at the same time you begin automating your initial processes so that you can expand your automation pipeline and should be executed by the project manager and automation leader. You will have to do this exercise until there is positive known traction in your company for RPA. This is a critical step and is often needed to continue to justify headcount and costs associated with the project to senior leadership or budget holders.
3.Beware of the idea of citizen developers. This person will look at automation as a bad topic unless educated first. Citizen developers should be something for mature companies to institute and realistically I question the viability due to the idea that low hanging fruits are typically automated first.
Second comes the more difficult processes which typically require logic and some additional coding skills specifically if you are going to do something like document capture. Some organizations are finding success by leveraging citizen developers (business process experts) to point to and identify processes that can be automated but take the automation only to a point where it is turned over to a developer or Center of Excellence (COE) to further develop and insure it adheres to the company’s security and governance framework.
There are additional topics such as how do you get to the processes via process mining and several softwares are being created to assist in this journey. This will be incredibly important after you have completed many of the obvious processes and you are on iteration 3 of process discovery. You can also use this software immediately depending on your RPA maturity and the willingness of the leadership to roll out newer technologies.