Reimagining life insurance underwriting

Life insurers have made steady progress over the past decade in automating their insurance underwriting. Platforms such as Swiss Re’s MagnumXP Assessment Engine can manage applications with speed and accuracy, delivering near-immediate coverage decisions for many people seeking the protection and financial security that life insurance brings. 

Even so, automation has its limits. Some applications that are initially assessed by automated platforms still require human underwriters to intervene, particularly when complex medical or financial risks are involved. In these cases, matching applicants with the right protection at the right premium depends on expert human judgment. 

To boost productivity in these instances, teams from Swiss Re's Group Digital & Technology Organisation and Life & Health Reinsurance worked together to build Magnum XP Underwriting Assistant. This platform utilises technologies including artificial intelligence (AI) to extract and organise key applicant data from referred cases, allowing underwriters to quickly apply their expertise where it is needed most. They avoid losing valuable time sorting through pages and pages of documents that may not be relevant. 

Underwriting Assistant summarises applicants' key data, adding clarity and consistency to insurers' risk selection while speeding up decision-making. While Swiss Re has deployed the platform in North America, additional proof-of-concept initiatives with clients in the UK, France, Germany and Asia demonstrate its global potential in lifting productivity of underwriting teams – whether they are facing capacity constraints, rising workloads, or simply aiming to make more effective use of their expert resources. 

The motivation to achieve fast, straight-through processing of life insurance applications through automated underwriting engines is clear: friction-free customer experiences can lead to more satisfied life insurance policyholders and higher policy placement rates, as it reduces the number of people who abandon the process because it takes too long.  

Globally, some markets see up to 90% of applications processed successfully via automated underwriting engines, though the average is around 75%. This is a strong result, but success can vary by region and by life insurance product. In the US, for instance, the most advanced automated underwriting engines still refer a significant share of complex applications to manual underwriting after flagging additional requirements that require human review. 

It's here where insurance underwriters become both detectives and historians. They must construct a data-based biography for each applicant, based on application disclosures, third-party data sources such as Medical Information Bureau (MIB), motor vehicle reports, drug prescription history, electronic health records, physician statements, physical exams, and lab reports, with the goal of making a coverage decision quickly.  

With risk-relevant data sources now proliferating, the challenge of making sense of all this information has intensified in recent years. Underwriting Assistant was created for just this moment, when insurers need support for their teams to boost the efficiency and accuracy of their risk-review processes.  

Using optical character recognition (OCR), natural language processing (NLP) and large language models (LLMs), Underwriting Assistant ingests unstructured and structured applicant data, normalises it to conform to common metrics, and organises it into a convenient, intuitive dashboard.

For instance, Swiss Re's platform highlights areas where the automated underwriting engine already had sufficient information about the applicant, while alerting underwriters to impairments where gaps still exist and require further expert review. This way, underwriters know immediately where to focus their attention.  

Some customers have described Underwriting Assistant as a "second brain" that helps expert underwriters tackle case complexity without experiencing cognitive overload. 

The benefits are manifold. We've observed a significant reduction in the time underwriters spend reviewing each case. Insights that Underwriting Assistant provides are also proactive, identifying anomalies in applications and proposing follow-up actions to quickly remedy information gaps so that an accurate, timely decision can be made.  

Clients using Underwriting Assistant also tell us that Swiss Re's platform helps them accelerate onboarding of their junior underwriters. This is due to its user-friendly dashboard with standardised views as well as the accompanying training modules that are well suited to support those who are still gaining industry experience.  

A more subtle but no less powerful advantage lies in Underwriting Assistant's feedback loop: the data it extracts from health records to help underwriters make decisions in complex cases can be used to inform models, helping refine an insurer's risk assessments over time. This can ultimately lift accuracy across their portfolio and improve an automated underwriting engine's straight-through processing rate. 

Underwriting Assistant works with any automated underwriting engine, including Swiss Re's MagnumXP Assessment Engine. It also ingests any underwriting referral rule system and integrates the insurer's underwriting manual of choice, such as Swiss Re's market-leading, AI-powered Life Guide. In addition, it supports multiple languages and can be used stand-alone or be integrated into a life insurer's underwriting workbench. 

Another client told us Underwriting Assistant had transformed their underwriters from hunter-gatherers who must painstakingly work their way through a dense thicket of data into teams of more efficient, productive risk decision-makers. That’s just the kind of progress that Swiss Re is aiming to achieve with AI: deploying promising technology to help insurance professionals get better at the things they do best.