Robots, biometrics and smart devices. If you had told me 10 years ago I’d not only have access to these innovative technologies but would use them daily, I’m not sure I would have believed you. Yet here I am, using facial recognition to unlock my phone, log into apps and access my bank accounts. And at Hyland, I am working with insurance organizations to understand, strategize and implement disruptive technologies, including robotic process automation (RPA), which works hand in glove with a content services platform. Much like a smart phone provides a variety of tools and capabilities to streamline our day-to-day lives, a content services platform and RPA can help insurance organizations improve their operations, drive efficiencies and meet their digital transformation goals — all of which helps them thrive in an evolving business climate.
The insurance industry has often been perceived as being slow adopters of technology, often relying on old processes and systems because they worked, even if inefficiently. What can we say…we’re risk-averse. Then, the insurtech era came along and presented us with different and more efficient ways of doing things. Keeping pace with all the new technologies can be stressful, but these technologies are impossible to ignore. As such, RPA has become a technology of interest because of the massive productivity and customer service benefits it offers.
In key business processes, getting all the necessary information and consolidating it often takes more time than deciding whether to issue the policy or pay the claim. RPA can do the gathering and consolidating without human intervention.
A human might grow tired of doing the same task of collecting and consolidating data again and again…that’s when things can be missed, and mistakes can be made. A bot doesn’t get tired. And, while the RPA bots handle more tedious, manual tasks, your staff is freed to focus on more creative work that drives customer satisfaction.
Setting an effective RPA strategy starts with structure
A successful RPA automation project relies on a vetted and structured format. To achieve this, organizations need to have control over the information feeding the RPA solution. Many insurers have implemented content services platforms as their information hubs, connecting all content within line-of-business systems and ensuring accurate, up-to-date content that RPA solutions rely on. Once fully connected, RPA and the content services platform provide a comprehensive suite to achieve intelligent automation.
To identify which internal processes, tasks and actions are the best candidates for automation via RPA, I recommend looking for those with the following qualities:
Standardization: Look for tasks that have a defined sequence and don’t have too much variance. Ideally, the work won’t change any time soon.
Structured data: The information and data that feeds the task should be relatively structured – or organized in a fairly predictable way so that it is easily classified.
Rule-based: The task or action should be built on a series of well-defined, objective rules. That means it would not require human interpretation to make a decision.
High-volume: The chosen tasks and actions should represent a substantial amount of staff time. Manually transferring data from one source to another is typically a good target.
Digital data: A task or action that already involves and relies on digital data is best suited for automation. If the task still relies on physical and handwritten documents, optical character recognition (OCR) and machine learning can be implemented to convert them to digital formats.
Implementation: Finding the right solution…and provider
How can insurers select a solution, and provider, that best fits their unique requirements? Look for an RPA solution that is scalable and configurable to ensure it meets your needs today and into the future. Selecting a solution that complements an existing content services platform or a vendor that can provide both creates an end-to-end, RPA-enhanced automation strategy — one that is designed to empower your organization to automate, optimize and transform tasks, actions and processes. Look for an RPA solution that helps your organization:
Analyze: Look for platforms that quickly, accurately and intuitively analyze tasks and processes down to the click level and automatically document process steps.
Build: The RPA solution should leverage a low-code toolset to allow you to quickly and easily create automation opportunities.
Run: Efficiency is key here – the solution should efficiently run unattended or attended automations, ensuring maximum bot utilization and scalability.
Manage: The best RPA applications manage and orchestrate bots with ease, using real-time dashboards for live monitoring and intuitive management.
When investing in any new technology, it’s also important to have a clear understanding of the total cost of ownership (TCO), which includes both the direct and indirect costs associated with the purchase. Be sure to calculate any additional fees for integrations, consulting, maintenance, training and other costs.
The current global health situation has led many insurers to accelerate their digital transformation strategies and new technology. RPA provides a great opportunity to enhance intelligent automation capabilities and further business process automation strategies. Insurers that leverage a digital workforce to complement their human employees provide employees with additional ways to excel at the work they do best, while delivering increased value for the organization.
In today’s data-driven world, bots help combat productivity drains that deplete resources and allow employees to focus their time on higher-priority tasks that build more meaningful connections with the customers they serve.
Artificial intelligence technologies are everywhere. The great leap forward in AI over the past decade has come along with an explosion of new tech companies, AI deployment across almost every industry sector and AI capabilities behind the scenes in billions of intelligent devices around the world. What does all of this mean for the personal lines insurance sector? SMA answers this question in a new research report, “AI in P&C Personal Lines: Insurer Progress, Plans, and Predictions.”
The first step toward answering this question is to understand that AI is a family of related technologies, each with its own potential uses and insurance implications. The key technologies relevant for P&C insurance are machine learning, computer vision, robotic process automation, user interaction technologies, natural language processing and voice technologies. It’s a challenge to sort through all these technologies, the insurtech and incumbent providers that offer AI-based solutions and where each insurer will benefit most from applying AI.
The overall value rankings indicate that user interaction technologies fueled by AI are at the top of the list for personal lines insurers. Every insurer has activity underway, mostly by leveraging chatbots for interactions with policyholders and agents or using machine learning for guided data collection during the application process. Insurers see high potential for transformation in policy servicing, billing and claims – areas where routine interactions can be automated.
Robotic process automation is in broad use across personal lines, although the RPA technology is viewed by many as more tactical. There is high value related to streamlining operations and reducing costs, but most wouldn’t put it in the innovative category.
Machine learning and computer vision have great potential for personal lines in both underwriting and claims. The combination of computer vision and ML technologies applied to aerial imagery is already becoming a common way to provide property characteristics and risk scores for underwriting. Likewise, images from satellites, fixed-wing aircraft and drones are frequently used for NATCAT situations. And AI technologies will be increasingly applied to these images for response planning.
There are many other examples. But for the purposes of this blog, the main question – which technologies are most valuable – has been answered. AI-based user interface (UI) technologies, machine learning (ML) and computer vision demonstrate the best combination of high value today and transformation potential for the long term.
But perhaps the more important question is not which technologies are valuable, but rather where AI technologies are most valuable in the enterprise. The short answer is that there are so many potential value levers and so many unique aspects to different business areas and lines of business that it is difficult to select just a couple of high-value areas. That said, it is relatively apparent that underwriting and claims both present major opportunities, and activities are already underway there. There are great possibilities for AI in inspections, property underwriting, triage, fraud, CAT management, automated damage assessment, predictive reserving and other specific areas.
There is no shortage of opportunities for AI in personal lines. Fortunately, there are increasing numbers of tech solutions in the market and growing expertise in the industry involving AI technologies and how to apply them. Ultimately, we expect to see a pervasive use of AI technologies throughout the insurance industry. Some will become table stakes. Others will define the winners in the new era of insurance.
With new technologies and evolving customer expectations driving rapid change in the insurance sector, research suggests that more than 65% of insurance carriers will adopt at least limited automation by 2024. But, today, the insurance sector largely relies on multiple layers of manual processes that make customers wait while employees try to make sense of complex documents.
Intelligent automation (IA) offers insurance businesses an opportunity to revolutionize the way they operate to meet increasing demands from customers and pressures from the market and to plan for future, unanticipated interruptions. Through the combination of robotic process automation (RPA) and machine learning (ML), IA solves complex enterprise issues through the end-to-end automation of a business process.
The Insurance Ecosystem Involves Many Parties and a Deluge of Data
Many third parties are involved in the end-to-end insurance lifecycle. That’s the case whether you are in commercial, employee benefits, retail or any other type of insurance. A lot of information gets passed around.
Brokers and insurers share data and documents. Advisers working with clients provide information, as do others, such as loss adjusters and lawyers. And the data arrives in the format preferred by the person who shares it.
That Data Comes in a Variety of Formats
Data used for insurance purposes comes in many formats — structured, unstructured and semi-structured — and must be ingested, understood and digitized with accuracy before any automated processing takes place. This involves making sense of data such as cursive handwriting, which is commonly found in life insurance change-of-address and name-change forms, as well as in beneficiary documents. Insurance entities must extract data from highly unstructured employee benefits documents, such as dental, income protection, long- and short-term disability and medical documents.
Brokers and insurers also need to compare and extract data from binders/slips, which can be up to 400 pages long and may use different words to describe the same thing. Insurers looking to ingest unstructured data (like email attachments, handwritten documents, PDFs and unlabeled data) — which is estimated to compose 80% of any enterprise’s data — can find their answer with cognitive machine reading (CMR).
While the industry’s standard data ingestion tool — optical character recognition (OCR) — can digitize structured data, it falls down when it comes to reading and extracting unstructured data such as tables, checkboxes and many other forms. In addition, OCR can’t read and digitize handwriting and signatures.
A CMR-enabled intelligent automation platform (IAP) can analyze and process large amounts of unstructured data and complex business contracts in a fraction of the time it takes with traditional, manual processes. An IAP enables insurance companies to address the error-, labor- and time-intensive challenges involved with human-driven processes.
For example, a global broking client wanted to extract 17 data points from commercial policies and endorsements. The documents came in from many different insurance carriers and in varying formats. All the data points required rules or reference tables to make the output usable, and most of the data didn’t have labels. In just a three-week period, with the samples of only 220 documents, with 40 different formats, multiple insurers and 10 coverage types, an IAP learned to extract 98.7% of the data, with 96.8% accuracy. Following this proof of concept, the client decided to implement this solution in multiple geographies.
John Hancock illustrates the many benefits businesses can derive from a CMR-enabled IAP. The company originally used manual processes to handle the large volume of policy management documents it received. Many of those documents held vast amounts of unstructured data — especially handwritten text in bold and cursive.
Since adopting AntWorks’ CMR-enabled IAP solution, John Hancock has enjoyed higher business productivity, lower turnaround times and a more than 65% increase in accuracy for handwritten cursive recognition. Because the AntWorks IAP uses assistive and adaptive machine learning to learn from exceptions, the system’s accuracy gets better over time.
Insurance entities also can greatly increase their case volumes with the help of CMR. Using manual processes would require armies of people to do validation checks and take a lot more time, while producing higher error rates.
One of the world’s largest human resources consulting firms implemented AntWorks technology to manage large volumes of data and provide optimized quotations to customers for new policies and renewals. This company eliminated manual keying and automated healthcare claims-related processes by extracting data from paper documents and validating for accuracy. That enabled 70% faster processing and a 40% increase in accuracy.
A Fortune 500 insurance company that provides title insurance protection and professional settlement services found that the manual process of validating title documents was leading to error-prone and inconsistent output. CMR technology enabled this company to increase field accuracy across orders by more than 75% and increase productivity by 200%. (Field accuracy is one of the key performance indicators that insurance companies, their technology suppliers and analyst firms like NelsonHall use to evaluate automation solutions. For example, NelsonHall in its SmartLabTest evaluation of document cognition platforms looks at the proportion of fields correctly recognized, accuracy of extraction of recognized fields and proportion of fields overall that are 100% accurate and require no manual intervention.)
IAP Equates to Faster Time to Revenue and Richer User and Employee Experiences
When insurers adopt automation, they dramatically improve the experience for all parties — the broker, the customer and the insurer. They relieve employees from doing what is considered value-added but repetitive work like manual data entry. Automation also eliminates the need for error-prone, stare-and-compare work that’s common in the insurance industry.
That elevates the customer experience because IA allows insurance companies to process requests and respond much more quickly. Digitizing processes also delivers a better experience because customers don’t have to contend with the cumbersome process of filling out and handling paper forms. Meanwhile, IA enables insurance businesses to enrich their data with both structured and unstructured data from other sources and use data analytics and predictive analytics to make their propositions better and more personalized.
IA also can enable businesses in the insurance ecosystem to move faster. That can help them to be more profitable.
Imagine a person is underwriting a life insurance case. If the data that person submits for the case is referred, an insurer would then have to go out to a doctor to get a medical report value. The underwriter would need to assess that report to understand whether it’s an acceptable case and communicate with the customer.
Getting and processing all the data can take weeks, delaying the policy kick-off. But if you can use intelligent automation to understand the data within medical reports, use rules to decide whether to accept or decline and automate the outcome, things happen much faster.
The title insurance protection and professional settlement services insurance company mentioned earlier reduced its processing time by 70% after adopting a CMR-based IAP solution. Meanwhile, the human resources consulting firm noted above increased its operational efficiency by speeding turn-around time, leading to higher customer renewals, an uptick in new customers and increased revenues.
Process Discovery Helps Companies Better Understand the Work They Do
Often, a lack of knowledge and understanding of process flows leads to automation failure. If you’re not quite sure which processes are the most optimal to automate or you’re not clear on all the steps involved in your process (and you don’t have time to do workshops with lots of analysts and business subject matter experts to figure things out), then process discovery is an excellent way of understanding exactly how the process is conducted.
Process discovery enables organizations to identify high-value processes for automation by recording actions that users undertake. If an organization can look at, say, 10 different people doing the same process, it can better understand not only how the process really works but also all the variations in the process, including things like the different process times and different applications accessed. The discovery enables the organization to see the steps involved and create automated processes that use the optimal approaches to those processes. The organizations can then apply what they learned to claims data extraction, fraud detection, mortgage verification and processing, account set-up, policy administration, quotation validation, title verification and a wide variety of other insurance use cases.
In addition to helping companies better understand their processes, process discovery can help in identifying opportunities for automation, expedite digital transformation and unveil previously unknown processes for in-depth process mapping.
Intelligent Automation Makes Companies More Resilient
Our new normal puts increased focus on the importance of business resiliency. Manual processes work against that because they often mean that workers need to go to physical business locations to handle paperwork. That creates risk in today’s environment. Intelligent automation frees people and organizations from on-site, paper-based manual processes and instead relies on processes that are better suited to today’s digital, distributed, remote work world. IA also scales, as needed, to adjust to changing circumstances.
The time has come for insurance companies to look at ways to improve their operational processes through technology innovation. IA has the capabilities to help insurance practitioners to do business much faster, more efficiently and with greater security.
In their search to boost profits and reduce their loss ratio, property and casualty (P&C) insurance carriers often turn to improving a cast of “usual suspects”: sales, pricing, new product development and a host of operational areas from new business through subrogation. But the biggest area to target— the one with the largest, near-term upside potential—is claims processing. Every insurer wants to reduce operating costs, cut claims leakage and reduce claim severity.
But what’s the best approach?
That depends on whom you ask. Technology providers insist that bleeding-edge, massive new systems are the answer. Internal processing teams will push for more human resources—with more relevant experience and better training. Other executivess will tell you to focus on reducing claims fraud.
But if you ask The Lab, we will say that the best approach is to keep asking questions, because the answers will point you toward a massive payback—a windfall. For example, what is the “standard” P&C claims leakage ratio, i.e., the industry average benchmark? And what is the source for this leakage number? Probably, the answer you get on leakage ratio will fall in the 2% to 4% range. Press for the source. The likely answer will be vague and hard to pin down. It’s unlikely that the answer will be: “our routine analysis and measurement of our claims processing operations—at the individual adjuster level.”
Stated differently, the source is actually “conventional industry wisdom.” If so, you’ve stumbled into a diamond field of improvement opportunities. To scoop them up, all you need to do is upgrade your company’s ability to perceive and manage claims processing at an unprecedented level of granular detail.
It’s worth the heavy investment of initially tedious effort. That’s because actual claims leakage is typically several multiples of this conventional-wisdom average of 2% to 4%: The Lab routinely documents 20% to 30%, and even more. That means that the payoff for reducing leakage, even for smaller P&C insurers, can easily reach hundreds of millions of dollars—which drop straight to the bottom line.
No, customer experience isn’t devastated. That’s because other (completely satisfied) policyholders are having their claims paid by adjusters who follow the carrier’s guidelines. The lower-performing adjusters, on the other hand, are simply not following these guidelines, and carriers fail to practice the process-management discipline necessary to ensure that all adjusters adhere to the loss-payment rules and targets.
Now, if you ask The Lab precisely how to reduce your claims leakage and loss ratio, we will point to three underused tools, or improvement approaches, to help P&C insurers surmount this challenge and achieve breakthrough levels of benefits, specifically:
Knowledge work standardization (KWS)
Business intelligence (BI)
Robotic process automation (RPA)
While the second two—BI and RPA—require a nominal amount of technology, the first approach, standardization, not only paves the way for the other two but also requires no new technology whatsoever. Typically, Knowledge Work Standardization, or KWS, alone delivers labor savings in excess of 20%, easily self-funding its own implementation— and readily covering much of the BI and RPA improvement costs. Taken together, these three tools rapidly transform an insurer’s P&C claims-processing operation and upgrade its related management capability. This allows management to significantly reduce loss payments while simultaneously improving operating efficiency. The result is an increase in “operating leverage”: the capability of a business to grow revenue faster than costs.
Interestingly, these three tools, or improvement approaches, also deliver major benefits for customer experience, or CX, aiding in policyholder satisfaction and retention. Here’s how:
First, roughly half of the hundreds of operational improvements identified during business process documentation will also deliver a direct benefit to policyholders.
Second, the process documentation and data analysis help pinpoint the reasons that policyholders leave. The predictive models that result help reduce customer erosion.
Third, these documentation and analytical tasks also identify the most advantageous opportunities for cross-selling and upselling. In this article, we will cover these three tools/improvement approaches broadly, then we’ll drill down to explore their real- world application—and benefits—in P&C claims processing.
1: The Search for Standardization in P&C Insurance Operations
Standardization—the same innovation that gave rise to the modern factory system—is arguably the most overlooked improvement tool in insurance operations today. And it applies to everything: data, processes, work activities, instructions, you name it. In other words, variance is standardization’s costly, inefficient evil twin. Consider:
Insurance operations performance is typically reported in the form of averages. These numbers are usually calculated for work teams or organizations. And this is also how supervisors approach their management task—by groups. Individuals’ performance is rarely measured, compared, benchmarked or managed.
Rules of thumb routinely apply. “Here’s how many claims an adjuster should be able to process in a given day or month.”
Industry lore trumps data-driven decision-making: “Claims processing is an art, not a science.” Or, even more dangerously: “Faster adjusters are the costliest ones, because they’ll always pay out too much.” (Spoiler alert: The opposite is true.)
Differences in details go unexploited: At one insurer, for example, The Lab discovered that five teams were processing claims—and each team used its own format and guidelines for notes. That single, simple issue confounded everyone downstream, as they struggled to reconcile who meant what.
“NIGO” prevails. The sheer opportunity cost of things like forms and fields submitted “not in good order,” or NIGO, can be staggering—often with tens of millions of dollars in unrecouped revenue flying just below executives’ radar.
2: Applying Business Intelligence, or BI, to Insurance Operations
Modern BI applications derive their power from their ability to create a clear picture from crushingly vast quantities of seemingly incompatible data. The best BI dashboards visualize this data as insightful, inarguable business-decision information, updated in real time. They let users zoom out or drill down easily; just think of Google maps. You can click from a state, to a city, to a house, then back up to a continent, using either a graphical map format or 3-D satellite photo.
Then why aren’t insurers routinely harnessing this power? Most already own one or more BI applications, yet they’re not delivering that critical Google-maps-style visualization and navigation capability.
This lack can be traced to two, intertwined obstacles: business data and business processes. Each requires its own, tediously mundane, routinely overlooked and massively valuable, non-technology solution: standardization.
Business datais already well defined—but it’s defined almost exclusively in IT terms. Think of the latitude/longitude coordinates on Google maps; do you ever actually use those? These existing IT definitions are difficult, if not impossible, to reliably link to business operations and thus produce useful, navigable business information.
The Lab solves this problem by mapping existing “core systems” data points to products, employees, transactions, cycle times, organizational groups and more. The solution requires standardizing the company organization chart, product names, error definitions and similar non-technology items. This is a tediously mundane task.
Technology can’t do this. But people can, in a few weeks if they have the right templates and experience.
Business processesare also already defined—but with wildly inconsistent scope. For example, the IT definition typically involves a “nano-scale” process—like a currency conversion or invoice reconciliation. Business definitions represent the polar opposite: global scale. Think of “order-to-cash” or “procure-to-pay.” All parties involved—throughout business and IT—thus talk past each other, assuming that everyone is on the same page. Worst of all, almost no business processes are documented. They exist informally as “tribal knowledge.”
The Lab solves this disconnect by mapping business processes, end-to-end at the same “activity” level of detail that manufacturers have perfected over the past century. Each activity is about two minutes in average duration. The range for all activities is wide but easily manageable: from a few seconds to five minutes. Over the past 25 years, The Lab has process-mapped every aspect of P&C insurance operations—and we’ve kept templates of every detail for these highly similar processes. Consequently, we can (and routinely do) map business processes remotely, via web conference… around the world!
Rigorously defining, standardizing, and linking business data and business processes underpins the best BI dashboards, delivering the Google-maps-style navigation that execs crave. This is how it’s possible to build astonishingly insightful BI dashboards that help make claims leakage losses apparent to our clients.
3: Robotic Process Automation, or RPA: A Powerful New Tool for P&C Carriers
Robotic process automation, or RPA, is simply software— offered by companies such as Automation Anywhere, Blue Prism, and UiPath—which can “sit at a computer” and mimic the actions of a human worker, such as clicking on windows, selecting text or data, copying and pasting and switching between applications. If you’ve ever seen an Excel macro at work, then you can appreciate RPA; it simply handles more chores and more systems. And it isn’t limited to a single application, like Excel. It is as free to navigate the IT ecosystem as any employee.
RPA “robots” are thus ideally suited for mundane yet important repetitious tasks that highly paid P&C knowledge workers hate to do. Better yet, robots work far faster than people, without getting tired, taking breaks or making mistakes. This frees up human workers for higher-value activities.
RPA also confers customer experience, or CX, benefits. With faster operations, customers enjoy the Amazon-style responsiveness they’ve come to expect from all businesses. On-hold times are reduced, claims get processed faster and the entire company appears more responsive.
Beyond the dual opportunities of knowledge-worker labor savings and CX lift, RPA holds the power to disrupt entire industries. Deployed creatively in massive waves, it can deliver windfall profits on a scale not even imagined by its purveyors.
Yet, today, most insurance companies’ RPA efforts, if any, are stalled at the very beginning; recent surveys indicate that internal teams hit a 10-bot barrier and struggle to find more opportunities, or “use cases.” That’s because the underlying processes to be automated are never made “robot-friendly” in the first place. So there needs to be scrutiny of the different activities—and the elimination of all of the wasteful ones that hide in plain sight, such as rework, return of NIGO input, and so on.
How to Overcome the 10-Bot Barrier in P&C Claims Processing
First, set expectations to focus on incremental automation with bots. No, you’re not going to replace an entire adjuster with a bot. But, yes, you will be able to quickly use a bot to call a manager’s attention to a high-payback intervention in the P&C claims-adjusting process. Examples:
Managers look for inactivity on open claims: If a claim is open with no activity in the last 10 days, that’s a red flag. But many claims are overlooked. A bot can call these out promptly.
Full-replacement cost, instead of partial replacement cost, is a major cause of overpayment that is most prevalent in roofing, flooring and cabinetry replacements. Bots can track payments and send management alerts based on line-of-coverage and even more granular detail. Roofing examples include: replacement, whole slope vs. whole roof; and number of roof squares replaced.
Audits are conducted on claims to improve quality and consistency—and to reduce overpayment. However, these are done on a very limited sample and only after claims have been paid and closed. Based on learnings from past audits, bots can alert management when certain claims- processing failures happen on a live basis. Managers can intervene… before payment.
Standardization (KWS), BI, & RPA: Focusing on P&C Claims Processing
All three of the above tools, or improvement approaches for P&C carriers—standardization or KWS; business intelligence, or BI; and robotic process automation, or RPA—can be readily applied to claims operations. Indeed, they seem to be custom- made for it.
Consider the following story, created from a mashup of different P&C insurance carrier clients of The Lab:
This “insurer” had plenty of claims data to share with The Lab; in fact, theirs was better than most. But that’s not saying too much: While 40% of the data was usable and comprehensible, the other 60% wasn’t. (Remember: This is better than most P&C insurers.)
Data was reported weekly, and sometimes daily, on an organization-wide basis. Here’s what they had data to report on:
Overall averages of claims processed, based on total headcount.
Average losses paid per claim.
That said, the company never tracked the performance of individual claims processors. They were all effectively “self- managed,” following their own individual procedures. There were no standard, activity-level instructions and guidelines, set by management, for quantifying targets for time, productivity or effectiveness. There were, on the other hand, vague, directional methods, many in the form of undocumented “tribal knowledge” and “rules of thumb.” The claims processors simply managed their own workdays, tasks and goals—similar to Victorian-era artisans, prior to the advent of the factory system.
When pressed, the company defended its choice to not track individual performance. The two reasons it gave would come back to bite the managers:
They were confident that individual performance, if measured, would only vary by about 5% to 10%, maybe 15% at most.
They were equally confident that imposing time and productivity quotas on processors would increase loss severity. In other words, they were completely sure that faster claims processing equates to overpayment of claims.
However, their very own data contradicted both of these notions—in a huge way:
First, the “long tail” of claims processors revealed a 250% variance between the top and bottom quartiles of individual performers—that’s 15 to 50 times higher than what management believed to be the case. In other words, the top three quartiles were out-processing the bottom quartile so much that there was no hope of the bottom quartile catching up—even getting close to the average. Put another way: Reducing this variance alone would yield a 25% capacity gain—an operating expense savings. And it could be accomplished by the top performers’ simply processing just one more claim per day—an increase they’d barely even notice.
Second—and just as important—the data revealed that the slower performers actually overpaid each claim by an average of 50%, an amount that totaled in the scores of millions, swamping the amount spent to pay their salaries. This carrier was thus getting the worst of both worlds with its lowest performers: They were slower, and vastly more costly. Not only that, they dragged down the average performance figures (not to mention morale) of the faster, leaner producers.
The impact from these revelations equated to losses measured in hundreds of millions of dollars. Incidentally, the story above is not rare; rather, it’s typical. As we’d mentioned, it’s based on a mashup of several insurance carriers.
Here’s one other standardization eye-opener. The claims process itself was rife with rework, turnaround, pushback and error correction. As a claim made its way through reporting, contact, dispatch, estimating, investigation and finally payment, it bounced and backtracked between the FNOL (first notice of loss) team, the appraiser, the casualty adjuster and so on. When presented with this “subway map” of the as-is process, the insurer’s executive sponsors were aghast:
Fortunately, the “spaghetti mess” can be cleaned up, even without new technology.
Business Intelligence (BI)
The Lab often encounters P&C insurance companies that invest heavily in systems such as Oracle Business Intelligence or Microsoft Power BI yet struggle to get value from these advanced analytics platforms.
Many of the issues stem from failing to “complete the final mile” when it comes to data definitions and hierarchies; that is, companies aren’t reconciling the IT-defined data elements with their own business-defined operations characteristics. This problem can often be traced to a disconnect between business leaders and IT organizations.
An IT person could—and often does—assemble and manage business intelligence for business units. But the person needs to understand the business so well that the person could confidently select which data to use and aggregate so that the final KPI (key performance indicator) in the resulting dashboard represents reality. And even if the person managed to create a BI picture of perfect “reality,” there’s no guarantee that the business would accept it. Let’s be frank: Creating useful BI and related analytics is a towering challenge. It’s overwhelming not only to IT; most businesspeople lack both the documentation and the comprehensive perspective to pull it off. So, the status quo continues: The “business language” experts will talk with the “IT language” experts, and the business executives will still lack the Google-maps insights they seek.
Another BI stumbling block is the “false precision” of too much data and too many categories. Consider the automotive insurer with “claims types gone wild”—such as “Accident: Right front fender,” and “Accident: Left front fender,” and so on. The Lab’s BI dashboards will often reveal to claims executives that 20% of the claims types represent 80% of the volume—another valuable, “long-tail” insight.
Robotic Process Automation (RBA)
As noted earlier, operational issues and customer-experience or CX challenges are typically two sides of the same coin. Often, both can be addressed by robots.
For example, consider the policyholder who calls the FNOL contact center and validates info. Then the person is handed off to another rep, who must re-validate the info. And then another. And another.
That’s not just an operational mess. It’s also creates a clear and present danger of losing that customer, hiding in plain sight.
While robots can speed repetitive chores, they can’t fix the underlying business processes (remember that FNOL spaghetti map, above?). Fortunately, Knowledge Work Standardization can. And once it does, the robotic possibilities are practically limitless: They span everything from sales prospecting to renewal notices to premiums/commissions reconciliation.
You saw how RPA bot deployments augmented the work of claims-processing managers. The next step is to augment the hands-on work of rank-and-file adjusters. Again, don’t try to replace the entire job position. Instead, augment the processor’s activities. In particular, hand off the adjuster’s mind-numbingly repetitive activities to the bot. This will allow the adjuster more time and thought—not to mention accountability—for complying with the policy’s payment guidelines.
For P&C claims, there are numerous opportunities to “park a bot” on top of routine, repetitive, knowledge-worker activity. Think of these as admin-assistant bots for adjusters. Here are two of many examples:
The “pre-adjudication assistant” bot. Adjusters spend lots of time sorting out “unstructured” information at the receipt of the FNOL. For example, they read descriptions of damage that arrive in free text data fields, then they standardize it and proceed to adjudication activities such as looking up coverages and setting reserves for the claim, prior to contacting the insured. Most, if not all, of these activities can be performed by RPA bots—but only if the inbound information is standardized. The Lab has used its KWS methods to create drop-down menus for this data and make it RPA-friendly. This standardization can be done incrementally, enabling bots to prep claims for adjusters: They look up coverage limits, set reserves and prep for the adjuster’s call to the insured.
The “customer contact assistant” bot. Adjusters, and others in the contact center, spend a great deal of avoidable and inefficient effort communicating with policyholders regarding their claims: advising status, notifying for damage inspections, obtaining corrections to initial NIGO information and more. Simply contacting customers can be a tedious, time-consuming and inefficient process; bots can help. They can be configured to send notifications to customers, preempting calls to the contact center. Bots can also initiate “text-call-text” notifications to customers’ cell phones. Here’s how it works: Bots, at the push of a button by the adjuster, send a text to the customer. The text may notify the customer to expect a call from the adjuster—avoiding call screening. The adjuster calls and gets through. Afterward, the bot sends a confirmation of the issue or next step.
Make the Move Toward Improved Insurance Operations & Reduced Loss Ratio
Claims processing, as we’d mentioned at the outset, is just one area within the P&C carrier organization where the power triumvirate of Knowledge Work Standardization (KWS), business intelligence (BI) and robotic process automation (RPA) rapidly deliver massive windfall value.
In the fast-moving world of insurtech, new technologies such as robotic process automation (RPA), intelligent automation, artificial intelligence (AI) and machine learning are making it easy for insurers to dream about transforming processes. However, too often they get lost on how to put those innovative technologies into use. For a risk-averse industry, we’re seeing more insurers open to using these modern technologies to improve processes and ultimately better serve their insured – the way they expect to be.
Many insurers are building on proven and integral technology platforms, including content services and core insurance platforms, to include more modern solutions that will help them further streamline operations. By combining capture, workflow, integrations and RPA, insurers can take some of the tedious tasks out of their employees’ workload and automate those processes, leveraging a “digital worker” to replicate redundant and manual tasks.
For example, take a loss-run request process – which one of our customers completely transformed using capture, workflow and RPA.
Intake processes are often tedious because there are too many manual steps. Without standardization around the process, it is inefficient and doesn’t provide reliable metrics. To continue to move critical information forward, data needs to quickly and accurately get to the right people – where and when they need it. Many content services applications offer multiple ways to capture data and instantly digitize documents, including emails, PDFs and Office documents, and connect them to key processes. This ensures data is digital from the beginning and throughout the lifecycle. Once imported and classified, insurers can create a standard way to kick off processes, drive additional efficiencies, enable performance metrics to identify trends and better assess internal resources. For loss run requests, once the request is made – whether by email or through the insurer’s portal – integrations with the content services platform can capture the request to officially initiate the process.
A workflow automation tool is an excellent way to help keep processes digital by electronically routing information to the appropriate person at the right step in the process. Additionally, because information is electronic, it is easier to monitor the status of items by incorporating real-time notifications. Within the loss-run request process, employees can use electronic workflows to take captured information and run that data through the applicable channels to get the claims history reports needed to make an informed underwriting decision. After the insurer receives those claims history reports, they can analyze how many claims were made, what types of claims were made and the financial impact of those claims.
RPA, intelligent automation, AI and machine learning are making it easier to take advantage of digital workers to further streamline processes and achieve greater efficiency. For the loss-run request process, once information is digitally captured, indexed and put through a workflow queue, the workflow can tell the digital worker to take indexed keywords and run them through third-party websites to gather any hits for the loss-run history. Once those are available, the digital worker can open the reports, download them from the website and upload them into the content services platform. There, the workflow process is finalized and the requestor, or agent, can access the report and make a decision. It took a digital worker three to five minutes to complete each item, saving more than 20 hours per day on run-loss automation requests, according to a customer using a combination of content services and RPA technology.
The entire loss-run request process was simplified down to nine steps:
Index data for transaction type, market and policy number in content services platform
Navigate to market web portal based on market keyword
Log into market portal
Navigate to “Request Loss Run”
Enter policy number and select submit
Retrieve loss run report
Save loss run report
Import loss run report to content services platform
Send report through final workflow steps
Building innovative solutions on proven technologies, like content services platforms, allows insurers to continue to evolve and modernize, as well as keep pace with the expectations of their clients. With any new solution, an organization needs to evaluate the best way to implement the technology into is business processes to ensure it helps them achieve greater efficiency and improve customer service. For RPA and intelligent automation it’s often easiest to incorporate and leverage these solutions in processes that:
Understanding the ins and outs of each of your processes is the first essential step to know where new technologies like RPA, intelligent automation, AI and machine learning will best benefit. Once implemented, employees become more productive and can focus on higher-value tasks to deliver faster and better service to their prospects and customers.