Streamlining Open Finance Integration
Financial institutions are challenged by the diverse APIs of open finance providers. Ninth Wave’s Compass system addresses this by normalizing bank APIs to the Financial Data Exchange (FDX) standard, creating a unified entry point for integration. This approach reduces the manual effort traditionally required for validation, mapping, and readiness scoring, a process that previously took weeks.
Leveraging Amazon Bedrock AgentCore
Compass utilizes a multi-agent architecture on Amazon Bedrock AgentCore. This architecture employs intent-based routing, where a central orchestrator classifies user requests and directs them to specialized agents. Each agent operates within a dedicated context window, preventing prompt space competition and improving accuracy. The system utilizes lightweight models for high-volume tasks and more powerful models for complex mapping and analysis.
Key Architectural Decisions
Three key design decisions underpin Compass’s architecture. First, intent-based routing ensures focused agent execution. Second, per-task model selection optimizes resource allocation based on task complexity. Finally, tenant-scoped grounding protects data isolation, ensuring that each bank’s data remains within its own session. This design leverages managed services like Amazon ECS on AWS Fargate, AWS Secrets Manager, and AWS IAM for security and scalability.
Enhanced Security and Compliance
The system incorporates enterprise security controls, including TLS 1.2+ and HSTS through Amazon CloudFront, AWS WAF, and robust identity management via OAuth2/OIDC with MFA. The architecture also adheres to SOC 2 and PCI DSS requirements, providing a secure and compliant environment for financial data integration.
