Data Analytics in Embedded Insurance Working Party


  1. Explore interesting open source embedded insurance data sets and discuss how data analytics techniques can be used for actuarial applications, including risk selection (for life and GI products), identifying insurance opportunities and market segmentation.
  2. Investigate the intersection of machine learning and traditional actuarial techniques in embedded insurance data, especially in the area of insurance product pricing and loss management


  1. Zack Chan (Chair)
  2. Xin Yung Lee (Deputy Chair)
  3. Niharika Anthwal
  4. Michael Gibson

What is Embedded Insurance?
The term Embedded Insurance refers to the act of introducing relevant insurance coverage or protection during a customer’s purchase journey of another product or service [1]. Customers here can refer to a provider and/or consumer of service of an online platform, such as a driver and a passenger. While other definitions of Embedded Insurance may exist, characteristics of embedded insurance common across these definitions include:

  • Use of personalised data to provide relevant insurance coverage and premium
  • Offering of insurance at a relevant point in time during a user’s purchase journey
  • Affordable and bite-sized premium
  • Fuss free underwriting and claims process

Data Analytics in Embedded Insurance
The emergence of Embedded Insurance is without a doubt the next evolution in the insurance industry. However, traditional actuarial analytics methodology such as pricing and risk management are still slow to adopt and leverage the large volume of alternative data collected. Examples include:

  • Using e-commerce shoppers’ purchase behaviour to estimate their lifestyle and needs to introduce and price relevant insurance
  • Adjust motor insurance premium loadings according to detailed driving habits collected through in-car sensors
  • Real time on-demand pricing of insurance risk based on drivers’ location and length of time spent on road