While approaches to commercial underwriting are changing rapidly, core systems are difficult to update quickly. Insurers are uncovering use cases for large language models (LLMs) and other forms of generative artificial intelligence (AI), but incorporating these emerging technologies into existing workflows can be complicated and full of risks that need to be identified and mitigated.
Fortunately, digital underwriting workflows are as unique as a carrier’s business goals, and the current insurance ecosystem offers no shortage of options when it comes to modernizing the systems and processes powering underwriting decisions. But knowing how exactly to determine the path forward can be a daunting task for insurance IT leaders.
Here are seven questions to consider when determining what type of digital underwriting solution your insurance organization needs.
- How many unstructured sources of data does an underwriter use today?
Underwriters typically look at several types of unstructured documents, including loss run reports, schedule of values, and inspection reports. Artificial intelligence (AI) capabilities are constantly improving, and insurers are starting to leverage AI to access relevant data and summarize risks.
Rather than replacing underwriters, AI solutions such as ingestion automation tools can be used to augment an insurer’s underwriters. Insights derived from tools like these, along with the necessary indexing and classification of received documents, can help expedite risk assessments and scale underwriting capacity.
For instance, carriers have been working to reduce their submission time by developing new business applications that cut down on manual data entry. These types of investments have sped up their submission rates by hours.
- What percentage of the complete view of risk is captured digitally today?
While several carriers have invested in application programming interfaces (APIs) and the like, a number of data points that underwriters refer to when making a decision are not captured or stored simply because the carrier’s systems do not have the means of recording the relevant information. The risk scores are saved, but the underlying raw data is often not available for future analytics.
When analyzed effectively, this underwriting data can prove invaluable in improving risk selection, and leveraging generative AI can further improve carriers’ ability to identify patterns. Generative AI can also be very useful for training new underwriters in best practices.
QBE North America leveraged an underwriting workbench platform to consolidate its underwriting data and provide a holistic view of its multi-line business in real time. Since implementing this platform and enhancing underwriters’ access to risk data in fall 2022, QBE reports improved underwriting accuracy and efficiency, with a projected 70-400% internal rate of return.
- Is the underwriting process optimized to focus on the desired market appetite?
Underwriters often know the type of market appetite that their organization wishes to write. The chief underwriting officer likely emphasizes consistent underwriting practices and ensures that some underwriting decisions are audited. Often this is done in response to a report that shows some deviation from desired characteristics.
An underwriting workbench that is properly enabled with analytics can help proactively validate the underwriting risk appetite. This type of intervention can improve the margins and overall profitability of the book of business and reduce deviations from guidelines.
One large property and casualty insurer ValueMomentum worked with wanted to increase its field agent quote conversion rate, for example, and leveraged analytics to predict each quote’s conversion propensity and prioritize those most likely to convert. Applying this analytical model helped the insurer improve its quote conversion and lift score as well as boost agent performance and satisfaction.
- How many days does it take to effectively onboard new underwriters?
Underwriting skill sets take time and patience to develop, largely because underwriting is seen as both an art and a science. But technology can help. The “science” component of underwriting can be trained more effectively using tools such as underwriting workbenches. Newer underwriters can be augmented with underwriting checklists, risk summaries, and recommendations.
In addition, the “art” of underwriting, which is most visible in nurturing agent and broker relationships or when sensing risk-related issues hidden within a submission (sometimes omitted from the record), can also be strengthened using specialized analytical capabilities (bind propensity) and LLMs (misclassification highlights).
Leveraging modern solutions effectively can prove invaluable in helping new underwriters achieve productivity more rapidly than with traditional methods of onboarding. Carriers such as Progressive, The Hartford, and Nationwide have leveraged customer analytics to guide underwriting decisions and new product creation based on behavior as well as economic changes in a given industry segment and demographic, while others, including Liberty Mutual and Allstate, are providing underwriters and agents with enhanced data access and faster processes through AI-powered tools.
- Can new data sources for risk assessment be added easily within the current process?
As risk assessment continues to evolve through advanced uses of data and the availability of such data increases, it is important to retain the ability to add new data sources within the underwriting process. Insurers are working not only with their own internal data but also with third-party data providers and government data sources like weather data from the National Oceanic and Atmospheric Administration (NOAA).
Having a process or workbench that is specifically designed to leverage such data sources and include them as appropriate is key to responding to market changes. When additional data is incorporated and analyzed effectively, carriers can see major results, especially as market factors and risk factors shift.
Utica National, for instance, was able to improve its classification and risk evaluation for small commercial policies by investing in a more comprehensive data solution. Working with more granular data allowed the insurer to achieve more accurate underwriting results while also automating portions of its underwriting process.
- Is the work process flexible enough to change routing as needed based on authority, referrals, analytical risk insights, capacity, etc.?
The pervasive use of data in generating risk insights and bringing underwriters up to speed quickly needs to be balanced with the flexibility to set up an organizational structure that can manage workloads effectively. Carriers engage with external vendors such as business process outsourcing partners, MGAs, or third-party administrators that might need information or direct referrals to underwriters.
To manage such flexibility and ensure appropriate access, the workflow needs to be set up in a way that organizes processes and people as well as leverages location for maximum effectiveness.
Carriers like Chubb and Travelers have started looking into ways that generative AI and other technologies can improve their workflows, including use cases specific to underwriting.
- Can the underwriting process be scaled to manage more distribution channels using existing systems?
Carriers are increasingly engaging with ecosystems of partners and expanding their distribution methods beyond the traditional channels. Commercial lines insurers are investing in evolving their agent relationships as well as boosting their direct distribution options, and specialty lines carriers may also offer program or wholesale business.
These channels can be profitable, but new methods of distribution mean an evolving workload for underwriters. And as workloads increase, underwriters can become a bottleneck. modern tools such as workbenches can help with directing case loads, leveraging partner-appropriate communication to complete pre-bind correspondence, and pushing some business through straight-through processing channels within appropriate thresholds.
Tokio Marine HCC – Cyber & Professional Lines Group (CPLG) tackled its ability to scale by combining its broker portal and partner API channels with its underwriting portal, resulting in a modern platform that helps brokers generate and bind quotes around the clock. This modern platform also helps CPLG respond rapidly to market shifts and necessary rating updates.
As the technology powering insurance workflows continues to evolve, carriers will continue to seek out ways to improve how they run their business and serve their customers. Underwriting is a critical part of the insurance life cycle, and modernizing the tools and solutions powering underwriting workflows will yield major results for carriers and their policyholders alike.
To learn more about how insurers are expanding their commercial lines success, read ValueMomentum’s case study NJM Grew Commercial Lines With Guidewire PolicyCenter Coverages.