1 Dec, 2020 • 3 min read

Leveraging Automation in Underwriting - Key Takeaways from our conversation with AXA, Zurich, Beazley and Swiss Re

In a recent Intelligent Insurer webinar, we gathered an all-star panel including leaders from AXA, Zurich, Beazley and Swiss Re to discuss leveraging automation in underwriting. Check out the key takeaways or watch the full conversation on-demand today.

Moderated by our co-founder and CEO Dr. Ed Challis the discussion focused on the opportunities and challenges of automation in the underwriting process. The all-star panel included

Key Takeaways

Uses and Benefits

Machine learning and AI are enabling new models for the calculation and automation of the operational processes around underwriting. Where currently underwriters are burdened with time-consuming manual tasks on often-archaic systems, these new technological models will increasingly drive efficiency in front- and back-end processes, improve customer experience, create more accurate quotes, and enable increased profitability without the need for continual employee expansion.

“It's something that we can't really afford to ignore” Tim Yorke, AXA

Automation improves underwriting in two major senses. It leads to faster underwriting through automated responses, the alleviation of time-consuming manual activities, and the standardization of processes. And to smarter underwriting via data-driven decision-making, cost-saving measures, and the freeing-up of underwriters to deploy their expertise where it is most effective.

“That's also a challenge, not only being fast, but helping our experts be experts instead of clerks.” Dr. Michael Zimmer, Zurich

These benefits synergise to create underwriting that is more efficient and effective on both portfolio and individual case levels.

 

Finding Balance

Underwriting can be incredibly complex. While the acceptance of risk cannot be fully automated, the surrounding processes can be. Automation will not replace underwriters; it will augment them with new tools, data, and support. Rather than substituting their roles, automation rebalances underwriters’ priorities from lower to higher value activities.

Underwriters, who would otherwise be treating cases in isolation, are empowered by new data streams that legacy operating systems lack. While underwriters will need to become more statistically literate, distinct roles remain for data-scientists and underwriters. Balancing the former’s new efficiency tools with the latter’s existing risk speciality will require close collaboration between the two roles while respecting the expertise of both.

“The ability to understand and interpret data is going to become more and more important for underwriters.” Tim Yorke, AXA

 

How Much, How Soon?

Low volume, high complexity products like health or commercial insurance should not be treated in the same way as transactional products. High-volume, low-complexity products, such as home and auto-insurance, are the most promising targets for automation. For some of these products, end-to-end automated risk-analysis and quote delivery is already possible in a matter of seconds. Additionally, the increasing use of data-capture technologies like remote sensors make aviation, agriculture, and shipping viable near-term options.

 

Other Important Takeaways:

Change Resistance - To combat organizational apprehension, start small and gradually integrate more machine learning, produce benefit-espousing educational materials, prove its long-term sustainability, and select ambassadors of change who will inspire confidence in the technology’s potential.

“People resist things they don't understand.” Tim Yorke, AXA

Data Aggregation Challenges – Scaling automation in certain insurance verticals will be difficult in geographies where publicly accessible data banks are unavailable.

Measuring Success – While top level KPIs will be expense-loss ratios, these will take time to come to fruition. Implement interim measurements to assess processes, underwriting efficiency, and the effectiveness of underwriters’ newly refocused priorities.

“How many more insights have I generated at a portfolio level? How many more strategic moves have I done? How many more impactful reallocations of resources and so on and so forth?” Claudio Morando, Swiss Re

Natural Language Processing (NLP) – Insurers have vast troves of correspondence and contractual data with which to implement many uses of NLP. These uses include sentiment analysis, contract wording, large-scale data mining, submission flow, the design of auto-triage and decline rules, and event detection that would allow for optimized pricing and claims handling.

“If anything, relationships are going to become more meaningful, I think they're going to become deeper.” Claudio Morando, Swiss Re

Remaining Personal – While low-relationship-value products such as auto-insurance are a natural fit for automation, the human element will remain vital in more complex products for the foreseeable future. Insurance is a people-focused business that benefits from trusted long-term relationships. NLP and machine learning will provide considerably more transparent interaction information with which to strengthen these bonds.

“With automation, with AI, we are able to bring it together and give the underwriter, claims handler, salesperson, the right information and with that information, he can be much more positive and ask the right question, and that's much more personal.” Dr. Michael Zimmer, Zurich

Outlook 2021 – Panellists identified the acceleration of automation as a key priority for 2021. Their focus will be on developing automation capabilities in the teams with the greatest potential for long-term growth.

Catch up with the full conversation ‘Leveraging Automation in Underwriting - today, available On-Demand at: today, available On-Demand. 

Blog-Inline-Intelligent-Insurer-Webinar

 

Resources Hub

The latest research, best practices and guides to delivering business value through Artificial Intelligence, Machine Learning, Data Science and Natural Language Processing.