As part of a major digital modernization initiative, Takaful Malaysia, a pioneering Islamic insurance provider, partnered with 3i Infotech to migrate its legacy insurance data into the PREMIA core system. The objective was to consolidate fragmented legacy data spread across multiple databases and Excel files into a unified Oracle 10g environment — ensuring data integrity, business rule alignment, and seamless front-end visibility.
This mission-critical project involved the transformation of over 1 million+ life and general insurance policies and more than 7.5 million related business entities, including claims, endorsements, certificates, customers, and accounting entries.
Takaful Malaysia’s data resided across three disjointed systems:
- TIS (Takaful Integrated System): Based on DB2
- FTS (Family Takaful System): Using Ingress
- Scattered Excel Sheets: Often stored locally by users
The target platform, PREMIA, enforced most business logic at the application layer — not the database. This introduced a significant risk: improperly transformed data, though present in the target database, would fail to surface on user screens due to compliance failures with front-end business rules.
To overcome this, the team needed a data migration strategy that validated legacy data not just technically, but also functionally — mimicking how the data would behave in the real application.
- Migrated 5 General + 3 Life insurance products
- Transferred 160,000 policies and 445,000+ business entities
- Migrated 18 Family/Life insurance products
- Transferred over 1 million policies and 7.5 million+ business records
- Defined Data Migration Scope: Finalized product list, modules, and constraints
- Source-Target Mapping: Matched legacy database fields and screen elements to PREMIA
- Data Profiling:
- System Profiling: Validated field types, lengths, mandatory rules
- Custom Profiling: Business rules (e.g., dummy Citizen IDs, payment modes, age limits)
- Data Cleansing: Removed polluted/invalid data based on profiling results
- Data Transformation: Handled data type mismatches and business rule gaps
- Staging Migration: Migrated ~2% sample to staging server
- Reconciliation Reports: Verified field-by-field data integrity
- User Acceptance Testing: Sample data validated by client
- Production Migration: Full data migration with pre- and post-migration validation
Traditional database-to-database migration was insufficient for PREMIA due to its front-end-driven validations. To handle this:
We introduced a third profiling layer — Product Profiling
Created custom scripts to push transformed data into PREMIA’s UI layer
Captured validation results (success/warnings/failures) in structured logs
Used the results to generate PREMIA Exception Reports for error resolution
This ensured business rules were enforced at the right layer — without duplicating the product’s functionality within the migration tool.
Identified common mismatches (e.g., date formats, null fields, expired policies)
Simulated real-time data entry into PREMIA UI to catch business-level violations
Broke down migration by business modules (Actuary, Claims, Underwriting, etc.) to reduce complexity
First from staging to production via UI; then synced database tables and primary key sequences
- Mismatch in Address Fields: Merged 4 source fields into 3 target fields
- Missing Gender Values: Derived from NRIC (odd = male, even = female)
- Missing Postal Codes: Filled manually based on customer address
- Different Payment Methods: Added missing methods to PREMIA pre-migration
- Encrypted Credit Card Data: Decrypted and re-encrypted with client’s IT team coordination
By combining deep insurance domain knowledge with innovative automation and validation techniques, 3i Infotech delivered a successful, high-integrity migration into PREMIA — setting the foundation for scalable digital operations at Takaful Malaysia.
Oracle 10g, UNIX, DB2, Ingress, Excel, PREMIA, Custom Data Profiling & Migration Tools
PREMIA is a product of 3i Infotech. This case study is an illustration of one successful project; results may vary based on individual organizational contexts.
