McConnell on Insurance: An RPA Reality Check

Squeezing efficiency out of decades-old legacy systems can seem like a losing battle for many insurance carriers, where long-term contracts are the nature of the business. Process engineers (whether in-house or contracted) have driven efficiency within these operations to the point of diminishing returns. After all, there is only so much that Excel macros and keyboard shortcuts can do to streamline highly manual and repetitive processes.

No wonder banks and insurance companies took note when Robotic Process Automation (RPA) entered the scene. By automating software application interactions — such as populating data, documenting audit trails and performing calculations — RPA promises to spare humans from performing these menial tasks while boosting accuracy, increasing efficiency and lowering the cost of operations by as much as 40%. Moreover, RPA requires minimal integration with legacy technology. By using RPA to streamline manually intensive operations, from underwriting new business to claims processing, process engineers can once again extend the life of these legacy systems.

Is RPA living up to the hype?

When properly applied, RPA has strong benefits. For example, at NTT DATA Services, we automated our proprietary life insurance BPaaS platform to bring greater efficiencies and a better user experience to our clients. By using RPA to automate task entry and acceptance as well as data extraction, our client, a global provider of employee benefit programs, saw a 99% improvement in quality and 78% improvement in efficiency.

However, many firms have struggled with deployment. One major cause is assuming that RPA could be deployed without relying on the IT organization. While this may be true for a proof of concept, scaling an RPA program so that a carrier can achieve real benefits requires the IT process disciplines that are core to the IT organization—such as setting up infrastructure, managing security and executing testing. Without this rigor, an RPA implementation may fail to deliver the promised results, and an otherwise promising robotics program could be scrapped.

The second major failure I’ve seen is managing the RPA program without an end-to-end view of the process, using experienced process engineers. Without this perspective, I have seen companies spend months creating bots that deliver little real efficiency because it optimizes only small portions of a process without re-engineering the overall process with the automation in place.

How to make RPA a success

A strong RPA program includes a cross-functional team that combines process and technology experts to re-engineer, develop and integrate with the people in the process. This approach:

• Maximizes the ability to identify the best places to apply robotics,
• Ensures that good IT discipline is applied; preventing rework,
• Implements re-designed processes with proper organizational change management to ensure acceptance.

What’s next?

Process mining tools are finding new utility as they integrate closely with RPA software companies. These tools allow for a more automated approach to opportunity identification and re-engineering. One tool that goes beyond process mining software is offered by FortressIQ. This company applies artificial intelligence to assist process engineers in assessing the processes being executed on the floor, and helps to automate the creation of the requirements documents; saving hours of effort and expediting the building of the bots.

RPA’s promise to enhance the customer and employee experience while improving the bottom line is boosting its adoption and spurring investment. In November 2018, RPA vendor Automation Anywhere announced a $300 million investment from the SoftBank Vision Fund. In September, UiPath raised $225 million to fund product development, deepen AI partnerships and invest in M&A initiatives.

Interest in RPA is booming. Forrester estimates that the adoption of RPA software will grow to $2.9 billion by 2021. Still, RPA is no silver bullet. When properly applied, RPA can drive efficiencies and improve quality, but it is an endpoint fix for legacy systems that are too difficult to change. RPA does little to improve the decision making that takes place within your processes.

In his blog post, Tanvir Khan, Divisional President, BPO, NTT DATA Services, observes that the marriage of AI to RPA will add the decision-making capabilities with RPA to drive additional opportunities for efficiency and improved customer experience across insurance processes. We see this combination for the intake processes at insurance companies. For example, using AI to improve Optical Character Recognition (OCR) and robotics to index and classify customer documents for proper routing within a workflow system is now becoming mature.

New RPA and AI applications will rapidly emerge with the help of knowledgeable process engineers. For example, detecting anomalies in data using AI can help find fraudulent transactions. RPA can then ensure that the relevant information is efficiently presented to the case manager for a thorough and efficient evaluation.

There is much to consider when implementing RPA, including involving cross-functional teams who will bring process reengineering expertise, IT rigor and change management know-how to the program. Has your company implemented RPA? What went wrong? What went right? Share your story with me via Twitter or LinkedIn.

Related Discussions

Learn more: Build a self-driving business that creates stronger customer connections and grows revenues

Blog post: Build a Self-Driving Business with SAP Leonardo

Blog post: Why RPA Will Not Be Any Faster Than the People It Replaces (and How AI Can Help)

Learn more: NTT DATA’s end-to-end solutions for insurance help you grow revenue, enhance ease of doing business and reduce operating/unit costs