Underwriting curiosity can take AI to the next level

Data and technology can be a game changer to better evaluate and manage complex and interconnected risks, but only when applied by human underwriters whose curiosity, sound judgment and emotional intelligence remain essential, writes Gianfranco Lot, CUO P&C Reinsurance at Swiss Re.

Data and technology have revolutionised the way we assess risk and make underwriting decisions. 

Today, underwriters can tap into a broader set of insights and information than in the past. With evolving technology like artificial intelligence, this will only grow in the coming years, as processing ubiquitous data from sensors, wearables, satellites and industry-specific data sources becomes second nature.  This can help them better understand, evaluate and price risk. Being able to properly leverage these types of insights can be a major advantage in creating robust and resilient portfolios.

The opportunities that technology and data present to find answers and get a deeper understanding of risk are exciting. However, a lingering question for me remains just what risk-relevant data do I need to help me underwrite, especially since there is so much data being produced and collected. To determine this, we need the human intelligence that underwriters bring to the table.  It's these human traits that are so critical: curiosity, the ability to ask the relevant questions and consider complex scenarios. We need people in the driver's seat guiding technology to areas that bring the highest value.

By pairing these human qualities with automation that technology enables, we have the ability to improve efficiency. For example, by using large language models insurers don't have to wait for claims to come in, they can instantly screen how many treaties will be triggered, their limits, potential waiting periods and types of exposure. Previously, this would have been a manual and lengthy process. Similarly for risk and treaty assessment, once an underwriter has defined the 'what-if' scenarios, AI-driven algorithms can help summarise them. That is very powerful, helping provide answers to evolving risks that we would have struggled to quantify even five years ago. These new capabilities inform underwriting decisions, portfolio steering and risk selection. They also open the door for us to develop new solutions.

We must avoid the temptation to rely solely on technology outputs. This is why it's vital that underwriters remain curious, critical and challenge and validate the outputs from AI-driven algorithms with their own experience. There must be a balance between artificial and human intelligence. 

Whilst technology can automate many routine administrative tasks and augment underwriters' ability to assess risks better, it remains a tool, not a silver bullet. AI hallucinations that fabricate information are a real problem that can deliver false information, potentially leading to poor underwriting decisions and liability claims. Equally, the issue of biases can lead to repetitive poor decisions. There are also other threats, like increased cyber risk. 

The re/insurance business, put simply, is often about investing in people's ideas. If for example a client wants to restructure a portfolio to improve performance or add a new distribution channel, human conversations need to happen first. Without the human factor, it is difficult to really understand intentions and understand how we can help clients through capacity or with other solutions.

The increased use of data and technology in underwriting is already reaping many additional benefits.  It is enhancing the re/insurance industry perspective, to one that enables more accurate use of forward-looking models rather than purely relying on historic data. For example, through CatNet®, our natural catastrophe risk assessment platform, we provide clients with AI and geodata enhanced tools to accurately forecast wildfire risks, amongst other perils. Further, in the claims space we're using AI-driven tools to provide granular, predictive natural catastrophe loss assessment data to clients. During Hurricane Ian, our Rapid Damage Assessment tool provided clients with insights into the insurance policies that could be triggered by property damage, helping them with reserving, planning loss adjustment and allocating claims resources. Technology is also leading to more informed discussions based on a better understanding of risks that can facilitate better ways to mitigate risk and develop innovative solutions. 

However, the biggest opportunity for us at Swiss Re is to use our aggregated global view on risk to deliver unique insights to our clients. This will enable them to either navigate risks better or identify areas to grow their business. The three questions I always ask commercial insurance clients is: do you know your large losses? Do you know whether you are under or overrepresented on a large loss? Do you know which large losses you were able to avoid? Given our access to large loss information, we can help clients to answer all the three questions. 

That's the power of data and claims information: to inform underwriting strategies, define risk appetites and help determine where adjustments are needed to stay competitive and profitable.

Data and technology excel in analysing vast datasets at speed and scale to find new insights that were previously inaccessible. Still, the magic only happens when AI is augmented with the curiosity and the experience of underwriters. It's therefore vital that we don't lose sight of the human element that is the backbone of the re/insurance industry, and which is needed to build loyalty and trusted partnerships. 

With conference season fast approaching, I look forward to sitting down face-to-face with clients and partners to understand your challenges and how we can help you either through capacity or through other technological and data-powered means that are guided by human experience.