By Stephen Elliott, MBA, JD, CISSP, CSM, SVP, IT innovation and decision optimization
Sometimes new technology comes along that does not directly replace old technology, but instead integrates with it and enhances it.
The claims process is still filled with legacy applications held together with manual tasks, so tremendous opportunity remains for automation without building entirely new systems. This is where robotic process automation (RPA) comes in.
Adjusters can spend quite a lot of time on activities that do not impact the overall outcome of a claim. Identifying low impact, repetitive and time-consuming tasks and moving them from a claims specialist to a robot that acts as a claims specialist is what RPA is all about. Robotic process automation is the use of specialized tools to “record” user actions and automate them using a bot that acts like a user. RPA records and plays back actions on a computer, like macros, but more powerful in that they can be used for any application. In certain cases, artificial intelligence (AI) can be leveraged to make different choices based upon the contents of the claim.
The benefits
RPA bridges technology — linking older systems to newer functionality in the claims process. Often a chatbot will collect data from a conversation, but no automated application programming interface (API) exists to enter that data into a back-end system. In addition, an examiner may be required to take some system actions based upon that data. In these scenarios, RPA can take the role of an examiner or an employee to enter that data, change claim statuses, enter notes or take other system actions that are not otherwise automated.
New claims intake processes are also perfect for RPA automation. And if single claim intake is strong, then spreadsheets filled with multiple claims are even stronger! Not limited to new data entry, an RPA bot can also receive changes to data and make those in a system — navigating screen-by-screen and making edits as required.
Where RPA really shines is when it crosses multiple systems. Data migrations between systems where an API does not exist are much quicker through RPA. Not to mention it avoids giving repetitive tasks to already overworked staff. Companies that acquire legacy systems with large amounts of data may turn to RPA to convert that data into their own system instead of attempting to create complicated APIs with a legacy system. An added benefit is that data validations and controls are often present already in the screens used for claim creation. The need to program those validations into an API can be avoided with the use of RPA.
RPA bots never sleep and can execute processes based upon events occurring 24/7/365. If events occur that need immediate attention — such as lockdown of specific accounts after a data breach or manual password reset processes when the help desk is overloaded — then RPA bots are available.
A bot can also monitor multiple systems. Whether it is checking dashboards or logging into systems and validating functionality on a periodic basis, RPA bots can be designed to notify administrators when something is wrong with the system that otherwise would be missed.
The challenges
As with all technology, RPA comes with its own set of challenges that makes it prudent to manage expectations of what RPA can and cannot solve. After all, the RPA bots still lack an examiner’s human touch and would not be the best solution for a call-back to an irate claimant. Also important, the rigor, controls and administration of an RPA system should resemble that of the implementation of a new programming language within IT or a new application. Version control, proper testing prior to putting a bot in production, regression testing after your system(s) change while a bot is using them are key concepts to the overall success and maintenance of an RPA solution.
This rigor — combined with the complexity of now having yet another IT system (RPA) running in the environment — may very well mean the costs outweigh the benefits of RPA. Finally, if your IT department built and maintains its own systems, it is entirely possible that what you want to automate with RPA can simply be programmed into the system directly — avoiding the need for an RPA bot to be engaged.
If used wisely and only on those scenarios where it is the best solution, RPA has shown itself to be a tremendous benefit for claims management. Faster data entry, automated processes, reduced costs and increased time for adjusters to focus on the people involved in the claims process are key drivers.