InsuranceERM Annual Awards 2024 - UK & Europe

Modelling solutions to empower insurance teams

Pierre Miehe, principal at Milliman Mind, explains how Milliman Mind is a portfolio of financial modelling and compliance solutions that offer insurers new levels of speed, agility and scalability

Can you give me an overview of Milliman Mind?

Pierre MieheMilliman Mind® is a portfolio of financial modelling and compliance solutions combining robust, secure and highly scalable technology with expert consulting. From regulation-specific tools to enterprise-wide modelling transformation, Milliman Mind empowers insurers' teams with new levels of speed, agility and scalability.

The solution aims to streamline existing processes, meet regulatory demands and drive modelling innovation – all while freeing employees from model management so they can focus on strategic outcomes.

The central component, Milliman Mind No Code, increases the productivity of actuaries and the quality of their models. It enables actuaries to create and manage a library of Mind "applets" – IT-quality applications for targeted solutions/processes.

Milliman Mind also includes Mind Streamline – a complete solution to streamline existing and cumbersome Excel/VBA processes, by converting them into a unified, secured and automated workflow-based process.

Milliman also has other solutions which are based on the Milliman Mind infrastructure:

  • Milliman Vega V6, the new generation solution for complete Solvency II reporting and XBRL generation ;
  • Milliman IFRS 17, the easy-to-use, quick-to-implement complete solution for IFRS 17 reporting.

Build production models/solutions/processes – there is no limit to the potential applications for Milliman Mind.

How can Milliman Mind benefit insurers? What are the technical advantages?

Some of the benefits of Milliman Mind are summarised below:

  1. Streamlined deployment of models: As a web- and cloud-based application, Milliman Mind requires no installation on local servers or workstations. Users can then deploy their models with just a "drag and drop", and set up a full workflow for their productions processes.
  2. Shorter time to production: Milliman Mind reduces the time and effort required to run models at scale — there is no need for insurers to rework their architecture or platform.
  3. Easy modification: Milliman Mind automatically converts Excel models to C# and builds a web applet on this basis. User can seamlessly convert models from Excel to Milliman Mind — and back again — with no functional degradation.
  4. Cloud-scale power: Powered by Microsoft Azure, Milliman Mind delivers high security, near limitless scalability, high performance, and easy maintenance and administration. There's no need for insurers to buy or maintain costly on-premises hardware.
  5. Governance and control: Milliman Mind enables users to control and restrict model access, with customisable rights management based on easy-to-configure roles.
  6. End-to-end auditing: Every change performed in Milliman Mind is tracked, whether it is on model inputs, model formulas or model runs, including when and by whom they were made, as well as who validated and signed off on the change.
  7. Full data analytics library: Building a secure AI-based model and putting it in production has never been easier. Milliman Mind enhances the full range of data analytics functions, like classification (including neural networks), regression (e.g. logistic), clustering (e.g. K-Means and K-Modes), distributions and copulas (e.g. Gaussian, Clayton, Frank etc.), hypothesis tests (e.g. Kolmogorov-Smirnov), kernel methods (e.g. probabilistic learning), and more!

What is the future of analytics in the insurance market? What factors will drive change?

Data analytics is progressively moving from classical generalised linear models to generalised additive models, and more advanced neural networks models. Also, in an increasingly competitive environment, price optimisation becomes a key differentiator to ensure insurers' profitability and is one of the main areas of research in the insurance industry.

But more recently, Large Language Models (LLM) are opening new horizons: more and more R&D is focused on what it could bring insurers, especially by boosting the productivity of data analysts. It might enable the testing and comparison of more and more machine learning methods, and may revolutionise the models used in insurance, just like AlphaZero from DeepMind (Google Group) revolutionised chess theory by making old IT principles outdated.

Within the Milliman Mind research centre, we have started an R&D project to investigate how LLMs could help our users get results even faster, and be more precise in their everyday work.

How intuitive is Milliman for the end-users, including both customers and internal staff? What level of customer support do you offer with Milliman Mind?

Half a day's training is enough to start benefitting from Milliman Mind's endless possibilities. We generally provide an initial 2-hours of training, and after one week, recommend a 1-hour dedicated session after users have started to experiment with our solution.

The idea of Milliman Mind is to make modelling simple. Simplicity is what will induce more productivity, and more control on model versioning.

A comprehensive online help, giving examples of each feature, is available for our users. We also have an online chat available for users to directly contact our team members.

Last, but not least, if insurers want to streamline existing complicated Excel/VBA processes, they can contact their local Milliman office who will offer a complete "packaged" consulting and software solution, and will assist in the development of a dedicated Milliman Mind process.