Interviews & Videos

Financial Strength Rankings using Artificial Intelligence

Interviews with Holger Bartel

In a series of interviews, Holger Bartel from the independent rating agency RealRate answers questions about disability insurance and the RealRate survey. All for the benefit of customers, because financial strength secures its future surplus participation:

Part 1:

What should be considered from the customer’s point of view with regard to disability insurance? What role does profit participation play?

Part 2:

Why is it important to pay attention to the financial strength of providers of disability insurance? Why are disability customers also affected by low interest rates?

Part 3:

The RealRate approach to rating insurance companies: Financial strength is brought to the fore. Using modern artificial intelligence methods, the causes of financial strength are analyzed and explained.

The insurance business is complex — let’s make it explainable

At the EAA Conference on May 12, 2022, European actuaries discussed Data Science & Data Ethics. Prof. Kraft (Coburg University) and Holger Bartel (RealRate Inc.) contributed with their novel approach to fair and explainable company ratings. We showcased our most recent rating of all German life insurers.

Our video on Explainable AI Insurance Ratings can be seen here!

Here you will find the scientific foundations on which the RealRate evaluation model is based. The paper „Causal Analysis with Neural Networks“ presents the artificial intelligence methods we have developed and used. The work on „Simple Solvency“ includes an early version of our expert system for the financial strength of German life insurers:

Explainable Artificial Intelligence (XAI) in Ratings

We will talk more about the RealRate approach that in the Meetup on October 28, 2021…

Explainable AI, Causal Graphs and an Application to Insurance Ratings

Holger Bartel, the founder of RealRate, shows how AI can be made explainable even for a small amount of data, like annual report data.

A structural neural network restricts the usually fully connected layers. Thus, the model is explainable by design. The methodology makes use of graph theory and derivatives.