Northern Trust and Microsoft announced an expanded collaboration on November 9 centered on an AI agent for institutional client onboarding. The application addresses a costly operational task: reconciling years of transaction and investment records into a standardized format before asset-servicing analytics can begin.
The partners report that the agent reduced data preparation time by 83%, from about two and a half weeks to three days, and achieved a 96.5% first-pass F1 quality score. Those figures describe their implementation and evaluation; they are not a guarantee of identical outcomes across all institutions, client portfolios or data formats.
The system uses Microsoft Azure and Foundry alongside Excel and SharePoint. Analysts review model outputs, resolve exceptions and approve reconciliations. Northern Trust says its client data remains governed under its control and that Microsoft engineering used anonymized information when improving the evaluation process.
The announcement is significant because regulated financial workflows demand accuracy, traceability and oversight, not simply fluent responses. Turning messy historical records into auditable structured data is a concrete enterprise application, but error handling and human supervision remain integral parts of the operating model.
The next question is how well the process generalizes across clients with unusual asset classes and inconsistent historical records. The release does not provide a randomized external study or a complete accounting of operating costs, so the reported gains should be treated as partner-measured results.