Improving the accuracy of ultimate net exposure analysis by 93% using actionable intelligence

Share

The client

We helped a leading global insurer accurately quantify their exposure to deadly Hurricane Laura within 24 hours of landfall.

The problem

When catastrophes strike, exposure managers are under intense pressure to quantify the ultimate exposure their businesses face. Traditionally, they rely on models and guesswork to do this given the lack of data from the ground soon after the event happens and the results are often inaccurate. This leads to poor use of working capital and a huge opportunity cost.

The Solution

A combination of machine learning and our expert analysts fused IoT data, weather data, scraped many open sources including social media and emergency services response, vessel and rig tracking, news and many other sources to generate analysis of the storm’s impact on the ground. This included an assessment of peril – it’s particularly important following hurricanes to know whether wind or flood might have damaged the insured location.

Our data science team then applied the exposure analysis to our client’s risk portfolio. It analysed their book to deliver an assessment of exposure against every risk location, and a highly accurate ultimate net exposure figure taking into account attachment points, layers of coverage and aggregation.

The Results

Our client was able to reduce their exposure reserving by over $100m from Hurricane Laura alone, and most importantly, proactively manage their claims process for policyholders at the time when they need it most.
Our intelligence delivers results that are proven to be up to 93% more accurate than relying on cat models.

Related