State Street deployed machine learning models in its middle- and back-office operations to automatically match and resolve trade reconciliation breaks, a task previously requiring significant manual review.

State Street's team of more than 400 data scientists, engineers, and analysts has deployed ML-based predictive reconciliation matching and break resolution in custody and accounting functions. The system also includes an ML predictive Net Asset Value (NAV) benchmark, client file quality controls, and tools for digitizing unstructured data from broker confirmation statements, trade instructions, bank loan agent notices, and funding memos. State Street publicly describes these tasks as previously manual or rules-based processes now handled or assisted by machine learning.