Insurance companies are embracing different machine learning techniques across a range of applications.
Most of the participating insurance companies apply different machine learning techniques, mainly including decision trees (e.g., random forests) and natural language processing (NLP) methods. However, other methods such as support vector machines (SVM), deep learning techniques (neural nets) and reinforcement learning are also implemented.
Interestingly, the most common applications are for text processing (58%) and model prediction (50%). Other use cases include outlier and pattern recognition (25%) and image recognition (25%). For the development of these applications, most insurance companies choose open source programming languages such as Python and R as their main development tools.
Insurance companies pursue various goals with the implementation of Data Science. One-third of all participants state that optimizing the customer targeting is their main goal in applying advanced analytics methods. Furthermore, 42% state that they aim for broader use of the models in the different business areas. However, insurance companies face several challenges when implementing data analytics methods. About half of all participants cited a lack of experience or expertise as one of the biggest challenges and 58% of all participants see integration into the existing IT landscape as one of the three biggest challenges.
EY’s uses analytical competency and business experience to support clients in the insurance sector. We help you connect the dots in the face of complex and disparate business issues, data and information. Using scenario modeling and analysis, we enhance the quality of data behind your decisions.
Summary
EY’s survey clearly shows that insurance companies are preparing for a data-driven future by investing in their data science capabilities, scaling up teams and testing new methodologies to enable customer engagement. However, insurance companies also face many challenges, such as a lack of experience, data availability and integration in the IT landscape. Choosing the right data and analytics strategy, governance and infrastructure is crucial for any insurer.