RPA and Cognitive Computing on the Verge of Fulfilling Promise in Insurance
When thinking about the demands of digitization faced by insurance companies today, it’s worth reflecting on the changes that mechanical automation brought to everyday life decades earlier — productivity gains, saved labor and increased comfort. That’s because we are on the verge of a second Industrial Revolution that promises similarly stunning advances that, like before, we will one day take for granted.
Robotic process automation (RPA) is rapidly maturing as a solution to drive vastly increased efficiency while simultaneously solving problems of systems discontinuity. The addition of cognitive computing on top of RPA, meanwhile, will introduce not only a greater magnitude of efficiency, but also the possibility of new applications that will revolutionize both back-office processes and customer service.
The RPA Opportunity
The contribution of RPA to business is revealed in the adjective it adds to the field of process automation: "robotic." The "robots" in question are intelligent systems — instances of software — that differ from earlier forms of process automation by their capacity for executing tasks traditionally performed by people. Like early manufacturing machinery, RPA is able to perform repetitive, rote, rules-driven tasks originally assigned to humans. Instead of moving physical objects, RPA executes tasks involving the manipulation of structured data, such as data entry, filling out electronic forms, and other information-related processes concerned with record-keeping and transaction processing.
Like labor-saving mechanical automation, RPA can drive significant improvements in efficiency for tasks that formerly required manual intervention. And while these tasks are nonphysical, they share the characteristics of machine-automated tasks in that they tend to be high-volume processes that are repetitive, well-defined and relatively simple.
They are, in short, the kinds of things that technology does better than humans. Machines can move much faster, for much longer and without making mistakes. In the case of IT, there's an added advantage of scaling at orders of magnitude higher than physical machines. And on top of it all, these benefits are location-independent. The cost savings remain the same regardless of whether or not the task is performed in a low-cost labor center.
What makes RPA particularly powerful for the insurance industry is that it addresses the chronic and widespread problem of systems discontinuity, an issue that necessitates manual workarounds. The completion of many administrative processes requires staff to manually transfer data from one system to another, potentially accessing multiple system interfaces for a given task. RPA enables staff to configure robots to access those systems themselves, connect them to one another, interpret relevant characteristics between them and execute tasks following the same rules that a human would use to do so.
When thinking about the potential of RPA for an insurance company, it's hard to overestimate the potential gains — in both efficiency and quality — from the application of the technology to the range of processes in a given enterprise that requires manual interventions between systems.
For example, DXC Technology and its subsidiary Xchanging perform claims services for various clients in the London insurance market. Seeking ways to improve efficiencies and lower costs, the team kicked off an initiative to focus on process automation for automating several key processes in the business.
The team kept in mind the types of well-defined, high-volume processes discussed above, seeking to gain significant efficiencies while affirming the principle that the initiative was not about replacing people with technology, but about continuous improvement. Not only is technology better at certain tasks, but these are tasks that people find unrewarding. Human capabilities, meanwhile, can be more effectively deployed elsewhere. In other words, RPA relieves people of the burden of administrative drudgery and frees them to do higher-value tasks — as in the case of expert systems, or workbenches for professionals such as underwriters.
DXC and Xchanging worked with RPA software leader Blue Prism to identify and automate 17 core processes. As featured in a case study by Leslie Willcocks, Mary Lacity and Andrew Craig, researchers at the London School of Economics and Political Science, one process identified for automation was the creation of London Premium Advice Notes (LPANs), documents that brokers use to submit premiums. The authors describe LPAN production as a high-volume process that operators strongly dislike, but which Xchanging has a contractual obligation to perform.
The process begins with the broker's sending of an unstructured data file, which must be opened and validated. Xchanging Claims Services staff must then add relevant data from a separate system. That enables the creation of the LPAN, to which the associate must attach supporting documentation and then upload the LPAN to a document repository.
The RPA initiative placed a robot at the point after which the broker submits the premium file. The file now gets handed over to a robot that has a distinct identity in the team as a task-performer. The robot has been "trained" to carry out various tasks that need to be performed. It reads the file and evaluates whether to complete it or treat it as an exception. Staff checks any exceptions, while the robot continues to create validated LPANs.
Staff worked with the robot to perfect its use, and within nine months, it had attained a 93 percent first-time completion rate. The LPAN creation process was reduced from several days to about 30 minutes, without error. The authors report that the robot also scales up and down easily to accommodate changing workloads, without any human resource issues.
The existing virtual workforce of 30 robots supports about 200,000 transactions per month and enabled the redeployment of full-time employees (FTEs) to higher-value work, in many cases working directly with clients. The initial return on investment was 200 percent. Today, the robots are maintained by two FTEs, who act as performance managers of the virtual workforce.
Beyond RPA to the Frontier of Service Automation
Because insurers have so many high-volume, low-complexity, rules-driven tasks, the opportunity to use RPA is clear. But for the same reason, the limits of RPA are equally clear. There are still many relatively simple tasks that are beyond RPA's ability to automate. Some fairly low-value tasks, for example, may seem ripe for RPA but utilize unstructured data in ways that aren't rules-driven or predictable enough. It is on that frontier that cognitive computing will be an enabler to unprecedented levels of service automation, reducing costs dramatically and helping clients respond aggressively to disruption in the insurance industry.
Cognitive computing — computer processes that mimic human intelligence using techniques such as data mining, pattern recognition, sentiment analysis and natural language processing — has virtually limitless application to business. Cognitive computing can improve back-office processes that rely on unstructured data or are more complex than RPA can handle. But where this emerging technology is likely to shine brightest in the shorter term is in customer-facing service applications.
DXC is proving this capability in the call center environment by deploying Watson Explorer, IBM's cognitive search and content analysis platform, as the backbone of a customer service dashboard. The interface integrates multiple systems and uses cognitive capabilities and RPA to spare the customer service representative (CSR) from having to execute multiple administrative tasks to get to the answer or transaction needed by the customer.
The typical insurer has acquired multiple back-end administration and customer file systems. Consequently, CSRs must interact with multiple systems to answer customer inquiries — which is time-consuming and tedious, and creates opportunities for human error.
Watson Explorer's role is to assist the CSR in handling the call by reducing related administrative tasks, but at a much more sophisticated level. This is possible because of the application's sophisticated tools and technologies, and because of its ability to handle unstructured as well as structured data across multiple back-end systems.
This leads to significantly reduced effort on the CSR's part in accessing and interpreting the information necessary to answer a caller's query or execute a relevant transition. These capabilities can significantly ease the CSR's burden and speed the time necessary to handle the call, but that is only the beginning of what the platform will be able to deliver.
DXC's insurance leadership team recognized the disruptive nature and new value from cognitive computing about 18 months ago and laid out a roadmap for delivering value for clients. That journey started with the implementation of an integrated dashboard for CSRs to help them provide more accurate answers faster. DXC is also now working with a team from IBM Watson to accelerate this journey with new cognitive capabilities to further assist agents and offer clients new options.
Ultimately, Watson's full range of capabilities could begin to have an impact on the customer interaction before the CSR even answers the call. The platform can recognize the caller and put the call into context with previous inquiries, for example, notifying the CSR that the person has called about the same issue recently or sometime in the past. Watson could subsequently follow the conversation through natural language processing and even apply sentiment analysis to gauge the urgency of the request or need to address the source of the customer's dissatisfaction. By "crawling" back-office systems, the platform can access necessary data and deliver scripting to assist the CSR in communicating with the caller.
Synergy: New Generation of Back-Office and Customer Service Process Automation
The synergy, then, of cognitive computing and RPA could increase efficiency by orders of magnitude. It can dramatically reduce the time needed to complete calls, and also multiply the proportion of inquiries that are solved to the customer's satisfaction on the first call. But its effects are more far-reaching. Since the system provides comprehensive support to the CSR, automating formerly manual processes, the time required to train CSRs can be greatly reduced — there's simply less for them to learn. The beauty of the blending of the technologies is that it reserves a space for CSRs to exercise their human qualities without being distracted by administrative tasks, resulting not only in much more efficient, but also higher quality customer interactions.
Insurance companies will soon begin gaining substantial efficiencies by applying RPA. It's an important part of the next wave of business process automation. The technology itself is mature and, through partnerships, companies such as DXC can bring the technology to bear in insurance-specific contexts. And as that work advances, DXC's partnership with IBM is shaping cognitive computing to meet the administrative and service challenges of insurance companies.
When combined, the synergies of RPA and cognitive computing will represent more than just an efficiency play. It's the beginning of a revolution in process automation that will inject new levels of efficiency and accuracy into back-office processes — creating a completely new generation both of self-service and high-touch human interaction.