Few businesses stand to gain as much from automation as insurance companies, but few businesses encounter as many challenges when they try to implement it. This is why we bonded together with our partner WorkFusion, a leading SaaS based automation solution platform that lets enterprise leaders digitize their global operations, to host “Beyond RPA: Optimizing standard RPA and the path to Intelligent Automation in insurance operations,” a webinar that focused on how the latest generation of RPA (which incorporates cognitive automation capabilities) can transform claim processes, an area that is too complex for traditional RPA. Our and Workfusion‘s goal was to encourage insurance carriers to compete in the digital era by giving them a playbook for how to fully exploit the capabilities of RPA and cognitive automation, so they can meet the increasing demands for faster cycle times, reduced expenses and improved customer satisfaction.
Not surprisingly, the topic really engaged our audience, who sent in many insightful questions that were thoughtfully answered by Milan Bhatt, Hexaware’s global head of healthcare and insurance and Anthony Russo, Director of Product Marketing, WorkFusion. These are the top five:
What is your point of view on touchless claims processing and how soon it will be adopted in the industry?
Right now, we think there are three distinct categories of insurers. The first one still works with a traditional approach with field investigations. Then you have the fast track approach, used by Esurance for example, which applies a mix of field and virtual touchpoints. Lastly, you have about 30-35% of insurance carriers who are actually using or moving towards touchless claims. We definitely see a movement there, especially in personal lines, and we believe that it’s going to be adopted by auto lines very soon. As AI and machine learning becomes more and more accurate, more and more customers will be open to adopting a process where you just take and upload a picture to report a claim and it gets handled entirely digitally, without any people involved. It’s a very fast process. We think that it’s still some time away from being adopted by commercial lines and workers compensation, but for other areas it’s around the corner.
What is the most successful approach to automate legacy processes on legacy systems?
We have seen customers take one of two paths when shifting to digital. The first approach, which is mainly used by personal insurance providers, is more disruptive. It entails investing a significant amount of resources to transform all of their core systems. The result is of course that all their processes are transformed, but they also gain the ability to change their product and make their distribution quicker, as well as transform their data. It takes a lot of effort, but the benefits are enormous.
For commercial insurance, we still see customers looking to adopt RPA to legacy systems due to budget constraints. We do believe that there several low-hanging fruits here, and that we can help customers achieve significant levels of automation with lower costs, but again, if you want to see radical results, it has to be a complete transformation where you start with core automation, move on to RPA and then machine learning once you’ve implemented the first two.
What are some of the best practices for insurance companies to consider while initiating a RPA pilot?
The approach that’s most likely to succeed is getting into a Minimum Viable Product or MVP mindset. This means setting a strategy to establish quick wins to build momentum and excitement at the executive level, and set up the implementation teams for success. We have outlined the following steps:
- First you have to settle on sourcing and implementation. Do you want to do your automation in-house, engage an incumbent vendor or outsource it? Then you need to choose development methodology. (We highly recommend agile).
- Your next focus will be process redesign and change management. This means selecting, prioritizing and optimizing the processes you want to automate. Standardization, although time-consuming, is necessary for ongoing maintenance and effective ROI. Change management is a central component of a successful IA initiative, and should consider vision, leadership engagement, communication, organization alignments, training and change readiness.
- Now it’s time to tackle technology and infrastructure. You need to evaluate the infrastructure readiness and compatibility and consider issues like bot ID, timeshare and hosting. Then you have to set up best practices for application development.
- In order to have a successful transition from analog to digital, it’s critical to not try to do everything at once. Scale your plan and continuously iterate on it. In order to have effective delivery you will need to set up a Center of Excellence (COE) with proper governance between IT and business.
How can you quantify the actual efficiency gain from implementing RPA? What are some of the parameters you should measure?
This is a question that often comes up, as ROI can seem kind of abstract at the start of a program. You will need to begin by defining success and know how to track against it. These are some factors that will help you measure benefits:
- FTE released from the process
- Number of transactions processed per day
- Time to process transaction
- Manual interventions reduced
How do I select and prioritize processes for RPA-led automation?
The most important first step is to redesign and optimize the existing business process. If you try to automate a process that is inefficient and disorganized, your automation process will be the same. Think of it as trying to build a structure on top of a broken foundation. Below, we have made a handy list of factors to consider when you audit and select your processes. All of these parameters may not apply to every process, they need to be discussed and selected on a case-to-case basis.
List of parameters to consider for RPA initiation:
- Front office process vs back office processes
- Critical v/s less critical process
- Degree of complexity
- Volume of transactions
- Repetitions involved in the process
- Number of systems involved to manage a process
- Manual hand-offs involved
- Amount of structured data input
- Level of unstructured data input
- Number of sources for data input, internal vs external sources
- Level of decision making/business rules involved (finite vs infinite rule scenario), complexity of rules
- Image processing needed? Variety of templates involved (finite v/s infinite)
- Need for natural language processing (NLP) (voice/chat) within the process
- Cycle time to complete the process
- Duration for which the process has to run
- Centralized (shared services) v/s de-centralized (at business unit level) process
- In-house v/s outsourced process
- Current state of automation / straight-through-processing achieved