How Munich Re HealthTech Uses Oracle AI to Power Insurance Analytics

Actuaries at health insurance companies work with large volumes of unstructured data from multiple sources to assess reserves, price products, and manage risk. For years, the process of gathering and analyzing this data was both thorough and time-consuming: assembling the figures needed for monthly reserve calculations alone took 10 to 15 days per cycle. Munich Re HealthTech’s SMAART platform was designed to compress that timeline, combining underwriting, pricing, claims handling, and risk management in a single application that could reduce reserve calculation time to 20 minutes.

As demand for SMAART expanded across multiple countries, two new constraints emerged. Insurers needed MRHT to address data residency requirements—regulations that mandate certain data be stored within the borders of the country where it originates. At the same time, customers were asking for more sophisticated analytics capabilities as their data volumes grew. Building those capabilities on the legacy Oracle Database infrastructure required significant development effort, and the team needed a path that did not involve rewriting its data model or application code.

MRHT migrated SMAART to Oracle Autonomous AI Database on Oracle Cloud Infrastructure, using Oracle Globally Distributed Autonomous Database to satisfy data residency requirements without architectural overhaul. The migration required no code modifications. Built-in database sharding policies automatically distributed data to OCI regional data centers in each customer’s country of choice, reducing global data replication and improving query latency. On top of this infrastructure, MRHT built an AI chatbot using OCI Generative AI, Oracle Autonomous Database’s Select AI and AI Vector Search features, and Oracle APEX for the front-end interface.

The chatbot serves actuaries, underwriters, and senior management, allowing them to query SMAART’s knowledge base using natural conversational language and receive answers in multiple languages. It accurately resolves more than 90% of the free-text questions posed to it. Oracle APEX also automated the extraction of data from spreadsheets, reducing the time to build analytical dashboards in SMAART from 15 days to approximately 20 minutes while eliminating human error. The project went live in four months, on budget, without requiring additional software licenses.

For the health insurance software industry, MRHT’s deployment illustrates how existing enterprise database infrastructure can be extended into an AI-native analytics layer without greenfield re-architecture. The combination of compliant data distribution, instant natural language querying, and automated reporting positions MRHT’s customers to operate with real-time insight into pricing, risk assessment, and portfolio management—capabilities that were previously accessible only through days of manual data assembly.