Northern Trust and Microsoft said on 9 November 2026 that a jointly built AI client onboarding agent has cut data preparation time by 83%, from two and a half weeks to three days, for institutional clients on Northern Trust Asset Servicing’s Front Office Solutions platform. Fintechbits covered Northern Trust’s last data deal in August 2026, when it plugged Lukka’s crypto data, drawn from more than 100 blockchains, into its asset servicing reports.

What the Northern Trust Numbers Measure

The agent converts years of a client’s historical transaction data, often including alternative assets, into Northern Trust’s standard format. Analysts review its work in Excel and SharePoint before a client goes live. The 83% figure comes from measured production cases against a prior baseline of about two and a half weeks.

The 96.5% F1 first-pass quality score is narrower than it looks. The footnotes say it comes from controlled testing across seven representative test cases, using anonymised historical client workbooks, before any analyst correction. Pete Cherecwich, Northern Trust’s chief operating officer, said the agent helps “get clients up and running more quickly during onboarding.” Faster onboarding matters for a custody bank because revenue only starts when the client goes live.

The governance design is the part other banks will copy. Microsoft’s engineers tuned the agent using only anonymised data, an “eyes off” approach that kept client data under Northern Trust’s control. Analysts stay in the loop to review outputs, check reasoning and resolve exceptions, and approved work feeds a governed knowledge layer on Azure. The bank cites its own survey finding that 57% of asset owners call consolidating data from many sources their biggest challenge. The firm had $19.9 trillion in assets under custody or administration at 30 September 2026, so even a modest gain in onboarding speed touches a lot of revenue. AI in onboarding cuts both ways, as industry leaders told Fintechbits when asked how deepfakes changed customer onboarding.