Enterprise Services Outlook Logo

Executing a Proof of Concept Test for Robotic Process Automation

By Terry E. Lang, Vice President, National Operations Manager, Comerica Bank

content-image

Terry E. Lang, Vice President, National Operations Manager, Comerica Bank

Savvy application purchasers in the contact (call) center space are starting to demand proof that their purchase will yield results before committing their hard fought technology budget. To this end, it is becoming more popular to run a proof of concept (POC) trial on the application and, if possible, run this trial as close to the real-time environment as possible. Robotic process automation (RPA) can have significant returns if implemented properly and is a prime target for an effective POC trial engagement.

The Institute for Robotic Process Automation and Artificial Intelligence describes RPA as the process of configuring “computer software or a ‘robot’ to capture and interpret existing applications for processing a transaction, manipulating data, triggering response and communicating with other digital systems.” Touting successes in lowering costs, higher efficiency, advanced analytics and greater quality, it’s no wonder RPA is popular for a number of back-office and in-line processes. In our example, the case was clear, RPA is a winner!

Start Small with the Right Use Case

Starting small and identifying the right test use case(s) will be critical in getting a quick sample and in building out an effective business case. The key for the test team lies in three components: What processes do we intimately understand, which of those do we have good measures for and which have ready process expertise? Controlling the test environment with the end measurement in mind will ultimately lead to greater success.

In our case we had a group of operational supervisors brainstorm eight potential processes and eventually settled on three: one that automated a manual process end-to-end, another which decreased in-/after-call work and finally, one that increased accuracy of duplicate entry. From these three business cases we knew many other production examples could fall in line.

Getting the Build Right

As discussed, what areas do you have command of and can clearly articulate the process from end-to-end? These are the best cases for a POC engagement. Who are your subject matter experts and do they have an interest in making their job easier? Do they understand the systems and all the steps involved in the current process? If these conditions are in place, the requirements gathering portion of an RPA test case can be completed in less than four hours.

"By Choosing Poc Cases That Have Existing, Complete Data Will Allow For Success “

For our engagement, we had our team leads engaged in ensuring the business process was mimicked in the RPA tool. These leads also conducted the pre-production testing and a 15-minute overview for their selected team members who would use the tool during our test period. Software coding was done by the perspective vendor using a no-cost demonstration license and a freeware optical character recognition engine for application entry.

Not all processes are good cases for RPA concept trials. Cases that involve subjective measures such as customer satisfaction, while an ultimate goal, do not lend themselves easily to POC case. Complex, inter or intradepartmental work handed off several times is also not a good candidate for the POC effort. Remember, if a solid return on investment can be demonstrated on a simple case, a larger and more complex work has the potential to yield far more rewards.

Monitoring and Measuring Results

The side-by-side comparison is the only way to measure the business case for RPA. By choosing POC cases that have existing, complete data will allow for success. What is the call or transaction volume on the selected process? Is there solid wage and system cost information readily available? Is the quality of the process easily measured and are variances easy to identify? Having a “before and after” measure in each of these areas can provide eye-opening results.

For our three POC cases mentioned earlier, there were some staggering results. The most impressive was a 30 percent reduction in call handle time for the in-/after-call work automated by the RPA application. Equally compelling was the ability to fully automate the entry of application data received in a formatted email and, in our quality case, improved our error rate to less than 1 percent.

How Far Can RPA Take You

The contact center industry is just beginning to see RPA simplify processes. Use cases are most obvious for back office processes such as application intake but are now being seen in such areas as using chat bots to automate interactive voice response interactions and automate simple transactions such as a balance lookup or account status. Knowledge base lookups exposed to customers, systems, and agents are other newer uses for RPA and artificial intelligence technologies. RPA automation is allowing contact centers to redeploy human capital to “true” customer service and complex sales transactions while accelerating customer satisfaction through easy to use robots.