Solution
Munich Re HealthTech migrated its SMAART insurance analytics platform to Oracle Autonomous AI Database with globally distributed sharding to meet multi-country data residency requirements. The cloud platform reduced data-extraction dashboard builds from 15 days to 20 minutes using Oracle APEX; built a generative-AI chatbot using OCI Generative AI and AI Vector Search that answers 90%+ of user questions in seconds; and enabled real-time semantic search across insurance underwriting, claims, pricing, and actuarial documentation.
Data flow
Insurance underwriting, claims, pricing, and risk data → Autonomous AI Database → distributed via sharding to regional centers; APEX dashboards extract and visualize data; AI chatbot receives natural-language queries → Vector Search → Generative AI → answers grounded in authoritative data.
Solution architecture
5 components · 2 layers - Storage
- Oracle Autonomous AI Database Core data platform for SMAART; stores unstructured insurance data, supports vector indexing, and enables natural language queries via AI Vector Search.
- Oracle Globally Distributed Database Provides transparent sharding for multi-country data residency compliance; distributes customer data across regional OCI data centers while maintaining single logical database.
- Serving
- Oracle AI Vector Search Semantic search engine for FAQs, documentation, and actuarial data; embedded in database; powers AI chatbot intent matching.
- OCI Generative AI Large language model service providing chatbot reasoning; answers 90%+ of free-text questions from business users with high accuracy.
- Oracle APEX Low-code development framework; reduced dashboard build time from 15 days to 20 minutes by automating data extraction and visualization.
- AI Vector Search inside the database for semantic retrieval
- OCI Generative AI for plain-language interface
- Oracle APEX low-code for rapid app delivery
- Oracle Autonomous AI Database with autoscaling
- Oracle Globally Distributed Database for geographic data partitioning
- Oracle Select AI for natural-language SQL
Munich Re HealthTech is improving customer engagement and reducing the time to create new offerings from 15 days to 20 minutes by powering its UnderwRiting Engine MAnager (UREMA) with Oracle Autonomous AI Database.
OCI Generative AI and the natural language interface in Oracle Autonomous AI Database, the team built a chatbot in days—achieving 90% accuracy and enabling underwriters to query data directly in plain language without relying on analysts.
PwC Germany leveraged the Oracle Database tools alongside its multi-year project experience to make the most of Oracle Autonomous AI Database and OCI Generative AI features.
