Building Financial Hierarchies with Amazon Neptune for Treasury Operations

This article describes how Amazon's Finance Technology team uses Amazon Neptune to model complex corporate treasury structures as property graphs for payment routing and cash management across hundreds of legal entities.

• Model financial hierarchies with labeled nodes (company codes, bank accounts) and directed edges (POBO, SWEEP) representing relationship types and transaction flows.
• Use Gremlin queries to resolve multi-hop payment routing paths, cash pooling sweeps, and bank account eligibility in real-time with single traversals.
• Implement Primary/State node versioning pattern where primary nodes hold current state and state nodes preserve historical versions for audit trails.
• Separate Neptune graph relationships from reference data stored in DynamoDB to minimize write pressure and maintain consistency.
• Iterate graph model incrementally from basic directed edges to labeled edges, bidirectional relationships, and node-level eligibility properties.
• Use distinct edge labels for different hierarchy types (POBO, SWEEP, legal entity) to support multiple coexisting hierarchies without query conflicts.
• Apply patterns to regulatory reporting, exposure aggregation, payment networks, and AML compliance across financial institutions.

Graph databases excel at modeling financial hierarchies through multi-hop, directional, rule-filtered traversals that support payment routing, cash pooling, and accounting path resolution at scale.