Northern Trust says Microsoft-driven data environments are cutting analyst prep time as it expands agent-based workflows

Northern Trust is leaning on Microsoft’s “FDEs” and an AI onboarding agent to speed up the way it builds and runs agentic workflows, the companies said in a new report. Northern Trust, which manages assets and provides trust and banking services, is using the approach to reduce the time analysts spend preparing data before agent-driven work can begin.

In the report, Northern Trust and Microsoft described an onboarding agent that helps accelerate setup for agents by automating parts of the early information-gathering process. The companies said that, in their tests, analysts saw an 83% cut in the time required for data preparation. The metric is framed as a time-savings improvement rather than a broader performance claim, such as accuracy or risk reduction, and it was not accompanied by disclosed sample sizes or benchmark methodology.

The term “agentic buildout” generally refers to the process of expanding systems that can take actions toward goals, often by coordinating tasks across tools and data sources. In this case, Northern Trust’s focus appears to be on building workflows where AI agents can be brought online more quickly, with less manual work spent on preparing the underlying data that agents need.

Northern Trust and Microsoft did not provide additional operational details in the reported account, including which specific agent capabilities were involved, how many internal teams adopted the approach, or how the workflows are supervised in production. They also did not specify whether the time savings were measured against a fully manual baseline, a partially automated baseline, or a prior iteration of an internal toolchain.

While the report is light on implementation specifics, the direction it points to matches a broader shift in financial services toward using AI not just for chat-style assistance, but for workflow automation that can span data ingestion, classification, and task execution. For an institution managing large amounts of structured and unstructured information, reducing the friction of data preparation can materially affect speed-to-deployment for AI initiatives.

Microsoft’s role, as described in the report, is tied to its foundation-style data environment components, referred to as “FDEs,” which the companies say help connect enterprise data to AI systems. The report does not break down how those components are configured for Northern Trust, but it suggests Northern Trust’s agent development cycle depends on having the right data ready for AI workflows with less manual staging.

There is also no disclosure in the report about cost impacts, model governance controls, data residency or encryption details, or whether the results generalize across asset classes, jurisdictions, or data sets. That means the 83% figure should be treated as a targeted improvement reported for the described use case rather than a universal promise.

Next, what will matter for investors and industry watchers is whether Northern Trust expands the approach beyond the initial onboarding scenario and whether it publishes more evidence around performance, controls, and operational metrics. If the time savings translate into faster deployment of new agent workflows, it could become a key internal lever for scaling AI in regulated environments, assuming the governance and audit trail requirements remain intact.