Description: Healthcare cost is increasing at 9.3% in Asia in 2017 and health insurance cost will eventually catch up, making healthcare unaffordable to aging populations. Inefficiency in healthcare system is a contributing factor. In this session, use cases on how AI is impacting the health insurance operations on 3 core challenges. Document processing: A lot of documents are generated in the healthcare ecosystem, medical reports, diagnosis, invoices etc. Currently most of the documents are still reviewed manually and rely heavily on users’ judgement. Customer engagement: Customers are generally under distress and feel rejected during the claim process. It is due to the time and process needed to get the information they look for, the gap between the sales and claims process, and the reimbursement amount. Claim settlements: Claim adjudication is a technical process that review the disease, treatment and policy coverage. The process is specific to disease and difficult to be generalized. Thus, the process is far from consistent and invites fraud, abuse and wastage in this area.

Description: PartnerRe Life and Health is rapidly growing its presence in global markets. This growth means that our medical underwriting team has to handle 25 percent more cases every year. To manage these additional cases without increasing headcount at the same pace, the team has to operate more and more efficiently. In a market where response time is a key differentiator, we improve efficiency by using machine learning to triage and prioritize the underwriting team's case work. In this talk, we describe the model pipeline end to end. That is, we cover the entire implementation chain from accessing data to the eventual evaluation of the added value. In particular, we discuss often neglected topics such as model risk mitigation, technical deployment of the model and integration into business processes.

Description: PartnerRe Life and Health is rapidly growing its presence in global markets. This growth means that our medical underwriting team has to handle 25 percent more cases every year. To manage these additional cases without increasing headcount at the same pace, the team has to operate more and more efficiently. In a market where response time is a key differentiator, we improve efficiency by using machine learning to triage and prioritize the underwriting team's case work. In this talk, we describe the model pipeline end to end. That is, we cover the entire implementation chain from accessing data to the eventual evaluation of the added value. In particular, we discuss often neglected topics such as model risk mitigation, technical deployment of the model and integration into business processes.

Description: Predictive modeling is a powerful toolset for assumption setting (mortality, lapse, etc.). The multivariate approach has its advantages in handling high-dimensional data, address correlations, and statistical significance. However, building insurance assumptions is far more than a simple modeling practice. One must take care to address how the model is applied and its business implications. This session will share the learnings from multiple projects.