Insurance Claims Platform Modernization

The insurer's claims platform had evolved over nearly two decades into a complex legacy system that was difficult to change. Core business rules were buried in stored procedures, while FNOL data from emails, PDFs and photos had to be entered manually.

The company wanted to introduce AI into claims operations but needed stronger controls around sensitive data, model access and compliance. Employees were also experimenting with public AI tools, creating additional governance risks.

CodeGeeks team used AI-assisted code analysis to extract legacy business rules into validated specifications and tests. The claims module was then gradually moved behind modern APIs, allowing the platform to evolve without a full replacement. In parallel, we built a governed AI layer that routes requests based on data sensitivity. Claimant PII stays within the insurer's EU environment, with PII redaction, role-based access, audit logging and prompt-injection testing.

We also delivered AI-assisted FNOL intake, automated claim summaries and a policy assistant with clause-level citations. Claims handlers retain final decision-making, while an approved internal AI assistant and usage guidelines provide a secure alternative to public AI tools.

Claims information can now be converted into structured records at intake, reducing manual data entry and speeding up initial processing. Legacy claims logic is documented and covered by tests, making future modernization safer, while the governed AI environment gives the insurer a defined and auditable way to introduce AI into daily operations.