Underwriting has traditionally evolved steadily. What does the current landscape look like compared to the past?
In my role as global Chief Underwriting & Claims Officer with a view across many markets, I have witnessed firsthand the rapid evolution of the underwriting landscape.
Following the introduction of underwriting automation well over a decade ago, which sparked a notable shift in underwriting practices, we have seen only a gradual evolution. Until recently.
Whilst I, and many in our profession, have been talking about the use of data, AI and technology for a long time, it’s only in recent years where we’ve observed a more significant acceleration in its material use within day-to-day underwriting.
Today’s underwriting environment is marked by a growing shift towards digital solutions and AI integration, which contrasts sharply with the practices of the past.
Traditionally, underwriting involved a largely manual process. After completing an application form, a tree-based rules engine would process these answers to make an automated decision for around 50-80% of customers. For the remainder, underwriting evidence would be required, and a human underwriter would make the final decision. I would previously spend my time considering how we grind out 1-2% increases in automation, marginally increase non-medical limits, or tweak application questions based on anecdotal feedback or lower quality data analysis.
Today, modern data-driven underwriting is increasingly characterised by its reliance on digital tools, AI, and alternative data sources, driving a more dynamic and streamlined process. This shift results in faster and more accurate decisions, ultimately benefiting consumers. Judgement and expertise are important, but now, when I’m considering any changes in underwriting practice, my starting point is always ‘show me the data’, before working with clients to better understand the right changes that will have the biggest impact on the most meaningful measures of success.
What are the most common questions you hear about the role of AI and technology in underwriting?
Two of the most common questions I encounter are:
- Can alternative technology and data sources be used to replace traditional underwriting approaches?
- Can (or will) AI or generative-AI replace underwriting?
These are also questions I ask myself and my team regularly, and my perspective on this has evolved in the last few years. But the short answer continues to remain “yes, but…”.
While alternative data, technologies, and AI have the potential to enhance and further automate underwriting processes, and in some situations completely revolutionise the way underwriting is carried out, I continue to see human underwriters playing important roles in this future.
However, in some situations, the concept of ’underwriting’ might not exist at all, as industries embrace the use of connected data to seamlessly assess new and existing customers frictionlessly. We are already seeing this to a limited extent today in many countries across Asia, with China, Korea, and India, leading the way on connected data and how this is used to assess underwriting risks.
For the majority though, and everything I work with clients on, the future is aligned with AI and advanced technologies augmenting traditional underwriting practices rather than replacing them entirely, with human oversight remaining essential for complex cases and nuanced decision-making.