Lessons Learned from our Robotic Employees

By Steve Leblond, VP of IS Operations & COO, Ochsner Health

RPA is an emerging technology that has received a lot of attention lately. A technology that automates manual functions currently conducted by employees; it was viewed as a solution that could reduce costs while delivering consistent results. After reviewing our organization found that RPA did have the potential to deliver significant value, and so like many others, we were excited by the possibilities. This was a capability that our Information Services division didn’t have initially but would benefit our organization if we did. As a result, we pursued a dual-path approach to leveraging RPA; first partnering with a vendor to move quickly, and secondly starting our internal program that could deliver automation as a new offering in our service portfolio.

Healthcare revenue cycle processes are often very manual and repetitive, as providing the appropriate information to scores of insurance agencies requires numerous people manually logging into insurance portals to upload data. As such, we saw this as a ripe target for automation and our partner was chosen due to their experience in this area and the speed at which they were able to engage. Learned two main lessons from our partnership efforts; outsourcing required more internal resources than expected and clarity on what systems you will allow your “robot” to write into is critical.

RPA is a useful tool now, but in the future, it will be coupled with AI providing further benefits.

Regarding the first, it probably should have been more obvious, but we believed that our outside partner would be very self-sufficient. It became evident quickly that our Information Services team needed to be closely involved since they proposed automated solutions. Automating a function exactly as a human didn’t always make the most sense. Humans find interesting workarounds to problems, and when you are programming a workflow, leveraging technologies that are being phased out isn’t always your best path. As our partner ran into the natural time, it takes for Information Services to review, understand, and execute their requests got larger to provide more flexibility and autonomy. Requests for 100 virtual machines so they would not have to ask every time they needed one does allow for more speed but at the cost of valuable resources that may sit idle for months or years. Virtual machines do cost money, after all, contrary to the perspective of many solution providers.

The second main lesson was that Information Services trusts our fellow employees more than we do our new RPA tool. A hospital’s EMR is at the center of all the care we provide. And having comfort with a computer accessing the EMR and potentially writing information into it was unnerving for some. The benefits of the solution suddenly became the problem. As an early adopter, our EMR manufacturer would not provide us much guidance, so it was “if it breaks something, don’t call us,” which is not a position with which any Information Services division is comfortable. And the fact that this technology can work quickly and without interruption gave us pause because if it does start incorrectly writing into the EMR, we weren’t sure how big the problem might get before it was identified. Couple these concerns with the fact that the largest cost savings items all point back to your EMR and you can see that it takes a lot of time just to get clarity and comfort…all while your RPA partner is waiting.

Our internally developed program has some lessons learned as well. Since it is an emerging technology, we did not have staff with any initial experience. We neither have an RPA tool, nor did we have a prime area to focus our automation toward as the ripest fruit was for our partner.

We set out to find our tool. When we started our selection process, we found that all RPA tools currently offer similar capabilities. Our initial thought of finding the “right” solution showed us that functionality wasn’t a differentiator. We quickly learned that while there was minimal differentiation between products, large differences often occurred in the manufacturer’s strategy and capabilities. We chose to go with a larger, pre-established company due to their advantage in human capital and their efforts in R&D. RPA is a useful tool now, but in the future, it will be coupled with AI providing further benefits.

Secondly, we needed to staff and train the team. We started with two resources, an engineer to do the programming and a business analyst to do process discovery and testing. We partnered with our RPA tool vendor for training during the first three workflows. Our approach was that the vendor would not touch the keyboard for those first three workflows, rather they would instruct our engineer step-by-step to expedite knowledge transfer. In discussion with the team afterward, they agreed that was helpful, but also acknowledged that they really did not have the full grasp of the tool until after they completed their 5th workflow, roughly ten months after the start of our RPA program. Our engineer in this space was a known high performer, so that length of time could be longer depending on talent.

Lastly, and probably most importantly, you need to ensure an effective process for reviewing proposed automation requests and rule out poor submissions. We found many poor candidates being submitted to our internal team. Some poor candidates were requests to automate work that creates little organizational value; 2 hours of manual effort that people didn’t want to do but would require weeks to automate. We got requests that require decisions that are non-standardized, a showstopper. And some poor candidates are requests that have no process defined at all. These made up about 40% of our requests. Our customers wanted us not only to automate existing processes but also to create and automate new efforts. Having the talent to create process was key. Our business analyst in this space was also a stellar employee with an excellent history in process creation, so we could successfully automate those as they often were value creation endeavors.

So as your organization pursues RPA, make sure you have the appropriate alignment and clarity on what and how it will interact with existing systems. And if you are creating your own program, remember that it is the people behind the tool that will make all the difference. While we love our new robotic employees, it is still our human ones that make them successful.


Share now: