January 17, 2019 | Written by: Andrea Eichhorn
Categorized: AI | Claims
Share this post:
Underwriting is at the very core of insurance. Depending upon the line of business, and the guidelines established by the carrier, the process can be very involved. The process requires a spectrum of skills, including a balance of art and science, along with lots of information, including structured—traditional quantified data—and increasingly unstructured—video voice, text, images, etc. At the same time that the complexity of risks grows, and the supporting information needing to be reviewed increases, underwriters are being asked to expand their roles across offerings to be more customer facing and, in many cases, sales-oriented.
New and emerging technologies increasingly support this ongoing evolution of underwriting. These capabilities are providing greater support and automation for data collection, compilation, and insight acceleration to free up an underwriter’s time to focus on making quality decisions more efficiently.
Where are the challenges?
In many conversations with underwriters over the years, I have found it helpful to deconstruct the underwriting process into a series of steps to help better understand the challenges at each of those steps. The six steps and some known challenges are described below.
1. Application submission: Receipt of submission and confirmation of the completeness of the application. Many submissions are received in paper, or images of paper documents, with a broad range of formats. Submission content may vary widely. Underwriting assistants often spend considerable time confirming completeness and following up on missing items.
2. Triage: assess and prioritize: Prioritization generally includes some combination of the following:
- appetite match
- conversion/hit ratio with similar business and/or similar sources of business
- carrier price competitiveness for the submitted risk
- quality of submission
- agency book of business experience
- experience with “risks like this”
- existing customer experience
- and more.
This step is a particular challenge to new underwriters, or those that manage across many lines of business, as they may not have a clear depth of experience with target metrics and, therefore, use their own best judgement in prioritization
3. Align content: augmentation and extraction: While this process may be iterative (starting in step 2), most carriers provide guidance to their underwriters on those areas to prioritize due diligence, and the depth of that diligence. Depending on the line of business/coverage type, this may be just a few quick verification of prior or related scores, entities, and any prior business, or may involve a much deeper understanding of the risks, evolving regulations, or other specific insights. We find that, on average, 50-80 percent of the underwriter’s time is spent on this work. This step also has the highest potential for inconsistencies as underwriters are balancing a range of priorities.
4. Evaluate: scoring and account rounding: Identification of the underwriting model and/or risk code (if used), requires judgement and experience, as does identifying opportunities to round out an account. Underwriters may not have full accountability for the offerings needed and may need to search their network to support relevant choices.
5. Decide: risk evaluation and quote: This step involves information to be input into the carrier’s policy system. This includes review of relevant coverages, exclusions to be considered, and the judgement of the underwriter. The output is a priced policy. For higher value policies, a quality review by a supervisor, or quality team may be involved. Having insights into how other similar risks may have been processed supports greater quality and potential revenue expansion.
6. Compliance and regulatory: Compliance requires understanding both internal processing policies, as well as regulatory requirements. These requirements or their application may change frequently and may vary from jurisdiction to jurisdiction. Adherence and auditing are critical for regulatory reporting.
Several of these steps may overlap, or in some cases be automated, to support straight through processing from submission to quote/bind and issue. Experience in underwriting is highly valued and takes many years to build. Yet the changing workforce is challenged by increasing waves of retirement being replaced by employees who are anxious to contribute, but may not yet have the experience needed. How do we bridge these challenges?
What are the on ramps?
Fortunately, many emerging capabilities can help carriers support the quality of outcomes of existing underwriters, while accelerating cycle time and concurrently ramping up new underwriter contribution levels more quickly. Often the challenge is on how to begin.
Carriers are assembling solutions that optimize their underwriting processes for differentiation and growth in the marketplace. Often this move may be as part of an evolving platform (like the IBM Insurance Platform), or one that is assembled by the insurance company for its own needs. Defining and implementing a clear operating model that shapes how these following three types of systems interact improves the overall agility of an insurer, while at the same time reducing legacy costs.
- Core systems: Modernize or replace with digital capabilities that can be accessed via microservices, and APIs to achieve flexibility and faster time to market. This can include a digital version developed by core systems providers, or a home-grown system that is developed on a foundation of microservices. These solutions generally include embedded predictive and cognitive models to support straight through processing, driving the outlier risks to underwriters for further analysis.
- Interaction systems: Use evolving technologies (AI, bots, voice, video, etc.) to create new ways of capturing information and interacting across key personas. Often a starting point is at the initial submission, using bots to more accurately capture information in a less “forms” oriented way, or in B2B, with the agent or broker, to support more meaningful interactions. These types of capabilities can be implemented fairly quickly (weeks or months) and deliver significant positive impact. These services are decoupled from the core system to support faster modernization while core system decisions are made and implemented.
- Insight systems: Draw information and insights from the ecosystem of core, and interaction systems, as well as select external sources, to drive processes, improve approaches, and drive improved quality. It is easier to start organizing and containerizing data from new sources; but ultimately the goal is to federate internal and external information in a governed manner. The focus of this path is to make it easy to access and provide tools for clickers and coders to create and utilize advanced models within the processes that they support.
Where are some starting points?
What are some of the ways we are seeing these ideas come to life in the underwriting process? Some examples include:
- Use conversational or text bots to interview prospects to capture information for applications. Then use process automation to input that information into the core system.
- Provide agents and brokers transparency to the underwriting process via a guided experience and supporting contextual knowledge base, and provide early qualification analytics.
- Use email automation to capture missing information.
- Create underwriting dashboards that aggregate information (internal and external) by risk type, by client, or by industry to reduce the amount of time underwriters spend aggregating information.
- Use analytics of “like this” risks to prioritize and help round out the offerings across a submission, improving cross sell.
- Use analytics to prioritize submissions that can be automated for straight-through processing. Use automation, predictive models, and rules to support the processing.
- Use baseline imaging via drones, sensors, and monitoring during underwriting. Use these baselines for risk reviews and during claims to understand changes in risk quality at renewal.
- Add external data to your process to anticipate risks. For example, use medical outcomes data and predictive analytics to assess current vs projected disease states.
- Provide simple apps to allow prospects, insureds brokers, and agents to check the status of an application. Allow permissions to add images, voice or content to their submission.
What will you do?
The possibilities for underwriting solutions are broad, and the impacts significant. Creative approaches like design thinking can help insurers prioritize starting points and ensure integration across the capabilities in all three areas: core, interaction, and insights. Decoupled architectures allow carriers to advance innovation more rapidly. Critical underwriting capabilities are accelerated and enhanced for greater value. The only question remaining is what is your path forward?
–>Learn about IBM Insurance solutions
–>Explore IBM Artificial Intelligence Consulting Services