POLICY ADMINISTRATION: RIPE FOR MODERNIZING
Advanced policy administration systems (PAS) are essential for insurers looking to compete effectively in today’s increasingly complex and dynamic market.
Ripe For Modernizing
Policy administration is at the heart of insurance operations. It provides the information that feeds core and secondary systems, such as document management and agent portals. Most insurers know modernizing these systems is an inevitable step on their journey forward, almost two-thirds insurers engaging in evaluating or implementing a PAS modernization initiative.
Many insurers are still experiencing challenges when configuring their systems. This ease of configurability is the number one feature insurers require in a new PAS, but is also one of the top challenges they cite in working with the solutions currently available on the market.
Evolving PAS solutions are adding and enhancing configuration capabilities, but their usability is still maturing.
Insurers are coming to realize they must reorganize to some degree to improve their maintenance processes. The skills and resources that a configuration specialist needs are often assumed to be present within the organization, but are sometimes lacking.
Resources that could be trained for this role are typically isolated within the business or IT organizations. Insurers must recognize that there are special skills that must be learned for PAS configuration and maintenance, and decisions to be made about where this work will be performed.
A recent study on Policy Administration: P&C Plans and Priorities, reveals that all insurers know they must take the modernization journey and that most are already on the way, although in different phases and at their own pace.
Fully 78% are planning to replace at least one core system (policy, billing or claims), and more than half are planning to replace all three.
Adopting certain best practices can aid insurers in working towards modernizing a company’s PAS, both before and after implementation.
There are challenges ahead, but they are not insurmountable, and the benefits of modernizing this core system will be felt for years to come.
Property & Casualty Insurance - Endorsement Solutions
A major focus area for PAS modernization is in the endorsement space. Taking an inside-out perspective towards these operations can help in transforming the overall architecture of the endorsement process.
Insurance endorsement is an additional document which is attached to the insurance policy that amends the policy in some way. Any insurance endorsement may add, delete, or amend any clause which is currently present on the insurance policy. Insurance agencies usually draft these endorsements on their own to set themselves apart from their competitors. Insurance endorsements are widely used in the industry, as they are readily available and are very convenient for insurers.
Amendments for property, liability, automobile or workers’ compensation and package insurance policies can be due to multiple reasons:
- Modifying property location
- Adding or deleting drivers or vehicles
- Payroll or scope of sale change for a business
- Mid-term clause addition such as subrogation waivers, additional interests, etc.
- Personal details of the insured change
The endorsement segment of underwriting support requires recognizing the changes being requested for various types of policies. This means teams must ensure the accuracy of sensitive data for each amendment while providing a high level of customer service. This can include activities such as:
- Adding prefixes to policy names
- Coding change requests
- Assigning specific individuals for individual policy types
Endorsement – Request Sources
A study of major commercial insurance businesses shows that these requests originate from the needs of the insured, but the majority of requests come from agents and brokers acting on their customers’ behalf.
The inputs for the actual process usually come through emails, faxed ACORD forms, and direct calls to field office agents.
Impact on Policy
Any changes to policies are mainly sustained by additional premiums or generating credits for the insured. Thus, endorsements are either premium bearing or non-premium bearing.
Automating the rating system for changes in premium can enhance the efficiency of this process. Additional interventions such as natural language processing (NLP), optical character recognition (OCR), intelligent character recognition (ICR), and artificial intelligence can also be used.
The endorsement request lifecycle has multiple hand offs with systems and processors across a single team. While insurers are often reluctant to move away from their legacy systems and processes, having too many applications and policy updates required when processing a request can lower the quality of results.
The validation of an email request is done only after the indexing, meaning a processor is responsible for vetting such requests and understanding if it pertains to correct workflow in the process.
Endorsement process - Policy administration scope
Since endorsements are also issued upon a premium audit of commercial insurance policy, it becomes imperative that the audit endorsement process is of high quality and on a fast timeline to provide better customer service.
Automating this process is possible, and can be streamlined by following several best practices.
Robotic process automation (RPA), OCR, ICR, chat bots, and NLP can generate significant improvements. A transformed customer-centric onboarding operation with improved human and machine interfaces can provide better customer insights, improved customer experience, increased revenue growth, and reduced risk.
- OCR/ICR can automate processing forms and mailroom operations, reducing rule-based manual tasks and data entry processes
- Using RPA for KYC checks can automate data collection and aggregation from various public and private sources
- Leveraging OCR/ICR for automated information extraction can significantly improve efficiency and accuracy when understanding customer requirements
Leveraging operations and automation capabilities can generate efficiencies through:
- Design thinking
- Structured intake
- Guidewire service platform
How BOT help read email requests with context
Enhancing the machine readability of documents to take into account their context and format, as well as considerations for regional slang, abbreviated words, or even hashtags efficiently requires a combination OCR tools, an NLP engine, and RPA. Such solutions can be quickly built and provide data accuracy rates from 75% to 90%, depending on the source,
For example, the syntactical changes between Arabic text in the UAE and Kuwait can be detected at accuracy rates around 78%. For English text across various cultures such as Australia, England, and United States, accuracy rates can reach 88%. Even when reading English from countries such as India and South Africa, accuracy rates can be 85%.
Intelligent automation also comes into play in cases with various data sources, as this requires determining the extraction technique for different sources and forms.
For example, say a brokerage company wishes to automate the process of collating and consolidating the documentation required for underwriting.
Typically, these companies have a document storage system for all the documents provided by a borrower. However, reading these documents and ensuring all the sources of funds are well understood and documented usually requires a human. However, intelligent automation solutions, which use a combination of OCR tools, an NLP engine, process automation, and machine learning, can enable straight-through automation.
Automation and Machine Learning for Underwriting
A typical solution for this documentation problem could be outlined in the following points:
- First go through the documents and identify bank statements
- Verify that these are all the bank statements as listed on the application form
- Identify all sources of funds. Typical sources of funds can include salary deposits, dividend income from investments, and stock transactions.
- A good intelligent automation solution should be able to correlate the salary deposits with any W-2 forms submitted by the borrower to validate the information
- Next, dividend income on the deposits should be correlated with investments declared and correlated with broker statements
- Similarly, credit reports from the IRS can be auto-entered into the system.
This solution enables the company to bring down its underwriting cycle from days to hours and minutes once the necessary documentation is available in the system. Such operational automation requires machine learning and the ability manage data from various sources.
Optical Character Recognition
OCR processes can enable organizations to improve their efficiency, accuracy, and reduce costs. This often takes place across several steps.
- Identify the type of the document, and scan an image with machine readable text into the system.
- Classify the document into understandable formats such as invoices, trade bills, or time sheets
- Read the document using OCR or ICR
- Interpret and derive conclusions based on text recognized on the document
- Act and assimilate, performing actions based on the conclusions such as setting up reminders, sending notifications, and storing data into a structured format.
Endorsements: OCR to ICR
Software can capture the image of a paper document, allowing the information to be translated to electronic data without manual input. The technology used to capture this image differs for each type of document.
Fixed-form and structured documents
- Claim forms
- Purchases orders
- Explanation of Benefits
Key focus areas of insurers
While many CIOs dream of providing their business counterparts with a system that improves product speed-to-market and delivers significant straight-through processing capabilities, these efforts are not easy. In fact, CIOs know that many significant and high-profile PAS replacement programs have ended in failure
Across the insurance technology landscape, there is simply nothing as daunting as a PAS replacement. While claims and billing replacements each have their own intricacies, they are relatively straightforward endeavors compared to policy administration. The technology ecosystem for policy administration is complex, playing a role in most of a carrier’s functions such as quoting and rating, to broader capabilities such as underwriting, distribution, and customer service.
In addition to the technical complexities of this highly integrated system, these programs become more difficult when attempting to streamline a carrier’s processes and improve performance at the same time. While a less complex “rip and replace” implementation approach is certainly viable, this strategy can be a long-term inhibitor of business benefits and does not allow carriers to drive additional value. Leading carriers understand that the real value lies in the ability to deliver improved business capabilities, such as improved product management, streamlined underwriting, and expanded distribution channels.
PAS implementations are costly and labor intensive as they utilize a great deal of both business and IT resources. Because even the most aggressive PAS implementations are measured in quarters, not months, there is a high probability that other transformation initiatives will be impacted. In addition to opportunity costs, the risks associated with these highly visible implementations are great.
But there are reasons to be optimistic about future PAS implementations. First, policy administration software vendors have greatly improved both their capabilities and the ease of implementation over the past decade. Second, using certain best practices can help insurers avoid common planning and delivery mistakes.
Simplified view of Policy Administration and its central positioning in insurance operations
Major drivers of change
In today’s competitive environment, carriers need a system that provides the flexibility to improve product speed-to-market, support the growth of its customer base, and improve underwriting capability and profitability. Customers and agents have higher expectations from carriers to provide improved customer experience and self-service capabilities. Reduced data entry and improved access to accurate policy data are important factors influencing producers in selecting their preferred insurance carrier.
Market forces have pushed insurance carriers to embark on PAS transformations to keep up with or leapfrog the competition. The specific functionalities that most carriers desire from their policy systems are:
- Simplified product development
- Quick quoting
- Improved straight-through processing from quoting and rating to policy issuance
- Enablement of new sales channels
- Quick and easy access to policy data for customers and agents
Many carriers are operating on systems that are decades old, which carry significant technology risks as well as increased costs to maintain and support. Older systems often fail to meet the changing demands of the business, due to their inability to scale fast enough to support the carrier’s growing customer base and expanding product set. Systems deteriorate over time due to the complexity introduced by multiple upgrades and enhancements over the years to support business needs. This makes them difficult to support and maintain. Maintenance becomes increasingly difficult as the workforce that supported these systems has retired or is close to retiring, and existing documentation is limited. In addition, the original software vendor has most likely terminated support for the system, requiring carriers to increase staff and associated operating costs to mitigate the risk.
IMPROVED VENDOR OFFERINGS
PAS vendor offerings have improved significantly in the last decade, offering carriers the flexibility and broad capabilities they require while limiting the need for customizations. Several vendors today offer products that support end-to-end insurance operations, including rating, underwriting, policy issuance, claims, and billing. A number of vendors also offer specific modules that are used to manage certain portions of the insurance policy lifecycle, and that also integrate with a central policy administration package. The increased number of stable and scalable vendor offerings, coupled with the increased appetite of insurance carriers to implement packaged policy administration with minimal configurations and customizations, has contributed greatly to the results of recent PAS implementations.
Drawbacks of transforming policy administration process
- Without clear business objectives, it becomes difficult to precisely define and prioritize a program’s scope and establish boundaries.
- The failure to lock down scope is one of the single biggest causes of PAS implementation failures.
- Without clear scope, requirements frequently change from a means of delivering the original objectives to a stakeholder wish list.
- Scope creep on any project is bad, but scope creep on any transformational effort this large and complex can be a disaster.
Why is transformation required in policy administration process?
Business objectives should be clear and tied to the business case. To accomplish this, consider limiting the delivery scope to a manageable level. The scope of the initial implementation should be limited to two of three major categories:
- Business process transformation
- New technology implementation
- New product introduction
Insurers typically recommend that a PAS transformation starts with the business process transformation and technology implementation. New product introduction should be a fast-follower in subsequent iteration.
What is particularly revolutionary about transformation software is that it does not necessarily require companies to make changes to their strategic processes or existing back office technologies. Even if companies are separated geographically or have various technological systems implemented, RPA is able to connect systems. Therefore, these tools may function as a low cost, low risk solution for process optimization with near-term payback.
Decrypting industry solutions
What qualifies for automation?
Despite the value potential RPA solutions can bring to the business, not all processes are candidates for automation. Advances in AI and cognitive computing are providing new opportunities to extend automation to processes requiring a high degree of decision making or involving a wide potential for exceptions.
Today’s RPA solutions are typically best suited for processes with the characteristics seen in the accompanying graphic.
Across the insurance value chain – including sales, underwriting, policy issuance, policy servicing and claims processing – there a number of manual and repetitive tasks that would benefit from automation. These include menial sub-processes where RPA could be applied to allow human employees to focus on the more complex, value-adding activities. Within the underwriting function, for example, the collection of policyholder records, prior year claims documents, and other underwriting data can be easily handled by RPA applications, improving the underwriter’s productivity and freeing up additional capacity to handle more cases.
This will not only lower costs, but also provide revenue growth opportunities by increasing the capacity to write more business.
Before launching an RPA project, a number of considerations must be taken into account to ensure that the full benefits of RPA are realized. First, RPA delivers the greatest level of efficiency when applied to processes that have already been optimized and where any non-value activities have been removed. Simply jumping straight into robotics exposes the organization to the risk of magnifying the pre-existing bad parts of a process.
Eliminate, optimize, automate, and robotics (EOAR), is one approach that can be used to prepare a process for the implementation of RPA. The first stage of this approach involves eliminating any unnecessary steps while optimizing the remaining actions in the process. The improved process is now ready for the application of RPA, first through the deployment of rules-based, process automation leveraging application performance management tools for unified monitoring and self-correction.
It is estimated that between 25%-40% of manual process steps can be automated in this phase. The use of analytics and cognitive computing provides significant opportunities to generate benefits through additional RPA implementations. The application of statistical analysis, prescriptive modeling, or workforce and workload monitoring can be used to automate an additional 10%-15% of processes within the business model. The execution of judgment-based processes, such as exception handling, is possible with the use of artificial intelligence and has the potential to increase automation by another 20%-25%.
Second, in addition to the RPA software and process improvement activities, a number of supporting components are required to successfully implement and manage RPA enabled processes
Automation can inject agility into the manual manuscript processes. Since most manuscript endorsement forms are not filed, they are not preprogrammed in the system or automatically picked up by the underwriting assistants at the time of policy issuance. Usually, the underwriting assistants draft these manuscript endorsements in a word processor and then paste them into blank forms that are preselected and presorted within the policy pack.
Copying and pasting content into the print system could cause multiple issues, such as loss of formatting, difficulty with comprehending images, and other issues. A more agile business process can alleviate this problem. If the business and compliance teams can draft these manuscript endorsements well in advance and deliver them to IT business analysts immediately after the account is bound, the IT team can potentially ensure that they are added into the system while the underwriting assistants register the policy.
As the underwriting assistant reaches the stage of selecting the forms from the system, the manuscript endorsement forms would be developed and readied to be picked up, making the entire process much more efficient and effective. Use data analytics to determine how many and what kind of manuscript endorsements were issued historically on accounts and draw similarities in their language for risks written within the same industry. For example, there many additional insured manuscript endorsements issued on a general liability policy.
Automation Process Case Studies
LEADING GLOBAL INSURER
BUSINESS PROCESS: POLICY RENEWAL PROCESS
- General liability and financial lines
- Non-standardized process between three different locations worldwide
- Process involved more than 25 applications and numerous documents, forms, and emails
- Part of the process involved an offshore business services provider
- Reduction of FTEs by approximately 50%
- 30 to 40% increase in efficiency
- Elimination of overtime, including peak cycles
- Increased customer satisfaction due to faster response times
- Increase in quality and accuracy output
GLOBAL INSURANCE BROKERAGE FIRM
ROBOTICS CENTER OF EXCELLENCE
- Identify business processes for automation
- Establish and demonstrate RPA capabilities within the context of the business
- Illustrate expected benefits and obtain business user acceptance on automated results
- 50% reduction in FTEs
- Reduced cycle time of targeted processes by 75%
- Reductions in error rates
- Ease of scalability to extend RPA
- Reduced operational cost