Gen Re Drives Efficiency and Innovation in Reinsurance with AI-Powered Automation through Posit Connect

Matthew Montero, Chief Data Officer at Gen Re, is responsible for the execution of their data and AI strategy. His team encompasses all IT-related data practices, including databases, data reporting, data engineering, data science, and AI initiatives.

When Matthew joined Gen Re, he observed that data practitioners across departments were operating independently and working with different forms of data, trying to create new solutions for the business. A notable challenge was the absence of a formal, centralized data science platform that offered unified security and data connectivity.

As a global organization, Gen Re must adhere to various laws and regulations, including GDPR, to protect personally identifiable information. Matthew’s goal was to put a system in place with pre-existing guardrails and security measures to enable data scientists to safely build whatever they want and focus on the data science work rather than work extraneous to that. This need ultimately led Gen Re to adopt Posit Team.

With the large number of actuaries at Gen Re, Posit was a clear choice. Many of their actuaries learned R as their underlying programming language during college and throughout their careers because of its statistical orientation. Coding offered them significant advantages over spreadsheets, such as reproducibility, transparency, auditability, and scalability.    

To support various regulations requiring data to reside and be processed in specific countries, Matthew’s team at Gen Re now hosts Posit Team globally, with 7 instances worldwide.

• Posit Workbench: 230 data scientists and actuaries across the globe (North America, Germany, Asia, Australia, and South Africa) can code in their development environment of choice — whether that is RStudio, VS Code, JupyterLab, or Positron.
• Posit Connect: Used to deploy dozens of applications, dashboards, APIs, and work products to make them available to a wider audience within the company and support business streams, whether that’s with the push of a button for simpler use cases or through a DevOps pipeline.
• Posit Package Manager: Custom packages are built and securely shared across Gen Re so that teams don’t have to reinvent the wheel for various projects.

Underwriting in insurance is often a time-consuming, manual process, requiring extensive review of numerous documents to make initial risk assessments. Continued improvements to this process are crucial, as they can not only lead to increased profitability with better risk management and less administrative burden, but also enhance customer experience by accelerating policy decisions. 

Chris Engelhardt, Senior Data and AI Operations Manager, leads a distributed team of data scientists, architects, and engineers responsible for developing and deploying custom AI applications and infrastructure across the Gen Re enterprise. One key initiative they launched in collaboration with other stakeholders is the Underwriting Early Indication project, which streamlines the preliminary risk assessment for primary insurers and offers early insights within existing underwriting workflows.

Gen Re leverages an automated process, powered by Posit Connect, to monitor and ingest new broker submissions every minute. These submissions are then sent to various AI services that extract relevant information. The workflow then provides a summary and an early indication of whether they should accept the risk. While human agency is still very much in focus, they’re able to accelerate the process significantly.

Gen Re utilizes Databricks primarily for data engineering, pulling data from various applications into their data lake and enriching it for business use. Posit, on the other hand, serves as the analytics platform for actuaries and data scientists, enabling them to build applications, packages, and dashboards using familiar languages like R and Python.

The two platforms work in conjunction as part of Gen Re’s enterprise data architecture to allow for managed and governed access to data in the data lake from a platform data scientists are familiar with — Posit Workbench. Additionally, for larger models and extensive computations, Gen Re can offload the compute to Databricks clusters.