北方信托与微软部署客户开户智能体：数据准备时间缩短 83%，由约两周半降至三天
美国伊利诺伊州芝加哥市与华盛顿州雷德蒙德市消息，北方信托（Northern Trust）与微软公司 11 月 9 日宣布扩大合作，推进北方信托全机构的人工智能转型，首项部署是为北方信托资产服务 Front Office Solutions 平台的客户上线一款新的人工智能客户开户智能体。
该智能体由北方信托与微软 Frontier Company 的前沿部署工程团队联合开发，处理的环节是机构客户开户中较复杂的一步：把客户多年积累的投资与交易历史数据，转换成可直接进入北方信托报告环境的标准格式。
平台的服务对象包括基金会、捐赠基金、养老金计划与大型家族办公室。北方信托称，开户时需要汇总多个来源的大量投资与交易数据以重建业绩历史，另类资产尤其如此。北方信托一项调查显示，57% 的受访资产所有者把整合来自多个不同内外部来源的数据列为最大挑战。
部署后的实测结果是，数据准备时间缩短 83%，由约两周半降至三天。智能体在受控测试中的首过质量得分（F1 值）为 96.5%，该指标综合精确率与召回率。微软在发布脚注中说明，F1 得分来自对北方信托提供的七个代表性匿名历史客户交易工作簿的受控测试，开户智能体的准备时间则在真实客户开户场景部署后测量，包含数据转换、分析师复核、异常处理与最终验证。
系统建立在 Microsoft Azure 之上，使用 Microsoft Foundry 的企业级人工智能工具，并与 Microsoft 365 中的 Excel 与 SharePoint 集成，供分析师复核、修订与批准智能体的产出。发布称，开发过程采用只用匿名数据调优的办法，北方信托客户数据全程留在该行自有管控之下，工作流各环节均设人工监督，分析师负责审阅产出、评估推理、处理异常并以专家反馈持续改进。
北方信托首席运营官 Pete Cherecwich 在发布中表示，该开户智能体把北方信托的资产服务专长与微软的人工智能能力结合，帮助客户在开户阶段更快投入运行，在保持客户预期的监督与管控的同时，减少准备与验证客户数据所需的时间。
微软全球金融服务企业副总裁 Bill Borden 在发布中表示，双方的合作展示了下一代企业运营模式的形态，即以人工智能放大人的专业能力、加快服务交付，并在设计中内置治理与人工监督。
北方信托在发布中给出的公司口径为：该行 1889 年创立于芝加哥，在美国 24 个州与华盛顿特区设有办公室，并在加拿大、欧洲、中东与亚太的 22 个地点设有机构。截至 2026 年 9 月 30 日，北方信托托管与管理资产为 19.9 万亿美元，资产管理规模为 2.0 万亿美元。
Northern Trust and Microsoft Deploy a Client Onboarding Agent That Cuts Data Preparation Time by 83 Percent, From Two and a Half Weeks to Three Days
Chicago, Illinois, and Redmond, Washington. Northern Trust and Microsoft Corp. announced on November 9 an expanded collaboration to accelerate artificial intelligence transformation across Northern Trust, beginning with the deployment of a new AI-powered client onboarding agent for clients on Northern Trust Asset Servicing's Front Office Solutions platform.
The agent was developed jointly by Northern Trust and Microsoft Frontier Company Forward Deployed Engineering teams. It handles one of the most complex steps in institutional client onboarding, transforming years of investment and transaction history into a standardized format ready for use within Northern Trust's reporting environment.
The platform serves institutional investors including foundations, endowments, pension schemes and large family offices. Northern Trust said onboarding can require consolidating large volumes of investment and transaction data from multiple sources to reconstruct performance histories, particularly for alternative assets. In a Northern Trust survey, 57 percent of asset owners cited consolidating data from multiple, disparate internal and external sources as their biggest challenge.
Following deployment, the agent cut data preparation time by 83 percent, from about two and a half weeks to three days. In controlled testing it recorded a 96.5 percent F1 first-pass quality score, a metric that combines precision and recall. Footnotes to the release state that the F1 score came from controlled testing across seven representative test cases using real, anonymized historical client transaction workbooks supplied by Northern Trust, and that the preparation time was measured in real client onboarding scenarios after deployment, including data transformation, analyst review, exception handling and final validation.
Built on Microsoft Azure, the agent uses Microsoft Foundry for enterprise AI and integrates with Microsoft Excel and SharePoint in Microsoft 365, so analysts can review, refine and approve its work. The release said Microsoft tuned the system using only anonymized data, so Northern Trust client data remained under the firm's control, and that human oversight is embedded throughout the workflow, with analysts reviewing outputs, evaluating reasoning, resolving exceptions and improving performance through expert feedback.
Pete Cherecwich, Chief Operating Officer at Northern Trust, said in the release that the agent combines Northern Trust's asset servicing expertise with Microsoft's AI capabilities to help clients get up and running more quickly during onboarding, reducing the time required to prepare and validate client data while maintaining the oversight and controls clients expect.
Bill Borden, Corporate Vice President for Worldwide Financial Services at Microsoft, said in the release that the collaboration shows what next-generation enterprise operating models can look like, with AI amplifying human expertise to accelerate service delivery and improve agility, with governance and human oversight built in.
Northern Trust said it was founded in Chicago in 1889 and has offices in 24 U.S. states and Washington, D.C., and across 22 locations in Canada, Europe, the Middle East and the Asia-Pacific region. As of September 30, 2026, Northern Trust had assets under custody and administration of 19.9 trillion U.S. dollars and assets under management of 2.0 trillion U.S. dollars.
发布：GSM·美国（发布链接）
图片来源：Wikipedia/Wikimedia Commons 词条主图：Northern Trust
GSM Global · 本篇英文版 English edition of this story
