The DiaLOG Chair is mainly focused on three distinct areas of study:
- Short-term horizon (1 year and a half approximately): study of the concept of customer value, in a context of digital transformation in the Life and non-Life insurance sector.
This axis focuses on two main themes, namely:
- The customer in the digital age: redefining customer value in the digital age, detecting and targeting profiles, optimizing product offers and improving certain transformation rates.
- Customer behavior in the digital age: acquisition of new customers, customer retention, study of termination and arbitration behaviors in life insurance.
Main research contribution: integration of a temporal dimension in the consideration of risk factors in statistical learning models.
- Medium-term horizon (about 3 years): improvement of technical risk management and management processes in Life and Non-Life insurance.
This axis proposes to study essentially:
- Pricing and provisioning issues in Life and non-Life insurance. In particular, it aims to enrich internal data with data external to the CNP in order to improve knowledge of risk, better anticipate risk excesses, and redefine best practices in terms of sharing vs segmentation.
- Optimization of claims treatments to improve customer satisfaction: services automatically triggered following a claim,...
Main research contribution: interpretability of ensemble models (bagging, boosting), study/definition of the importance of risk factors on the phenomena studied.
- Long-term horizon (5 years approximately): study of the future impacts related to the evolution of environmental factors in Life and Non-Life insurance.
It is a question of thinking in order to acquire a prospective vision on:
- The impact of the extension of life in life insurance (mainly dependency risk), and of the delay in retirement (work stoppage, psychological).
- The use of AI to measure the impact of environmental risks (on insurance claims and on health): climate change, pollution,...
Main Research Contribution: taking into account outlier data, both in machine learning algorithms (missing data, extreme data,...) and in more traditional actuarial methods (GLM, GAM,...).
All the themes mentioned above can continue to be studied throughout the duration of the chair, regardless of the horizon previously established.
