JPMorgan Chase is placing more weight on redesigning workflows and strengthening control systems than on automating individual tasks as it expands AI adoption. [Photo: Shutterstock]

JPMorgan Chase, the largest U.S. bank, is redesigning work procedures across the board to increase the benefits of adopting AI. The bank judges that even if AI processes individual tasks quickly, it is difficult to achieve tangible results if overall workflows do not improve.

American Banker reported on Oct. 20 that Gill Haus (길 하우스), chief information officer for JPMorgan Chase’s Consumer and Community Banking, stressed in a recent interview the importance of measuring productivity, redesigning work and strengthening control systems following AI adoption.

Software development is a key example. Even if AI speeds up coding, if testing, documentation, security reviews and deployment do not move at the same pace, bottlenecks simply shift to other stages, he explained. JPMorgan Chase is expanding its use of AI from coding to writing deployment documents, analysing anonymised operational logs and supporting system outage resolution.

Haus said, "We do not hire engineers to write code, but to decide what code should be written."

The bank is also re-examining existing automation approaches. Rather than simply replacing robotic process automation, or RPA, which carried out repetitive work between unconnected systems, with AI agents, it is looking at modernising legacy systems to eliminate unnecessary work itself.

It is also important to confirm whether AI contributes to actual business performance. The view is that companies should compare processing time, errors and resources used before and after adoption, and assess whether it led to cost reductions or service improvements.

The authority granted to AI is also a key issue. JPMorgan Chase is blocking external AI agents, including Meta’s Muse, from unauthorised access to customer accounts.

Jamie Dimon (제이미 다이먼), JPMorgan Chase’s chief executive officer, warned during an earnings announcement on Oct. 13 that "uncontrolled AI agents could pose risks to customers and the bank."

Haus explained, "Unlike existing RPA, it is difficult to fully predict the actions of AI agents." He stressed that companies need reliable data, access-rights management, ongoing monitoring and procedures for intervention by human supervisors.

JPMorgan Chase is focusing on redefining the roles of people and AI and proving the effects, rather than simply expanding automation.

Keyword

#JPMorgan Chase #Gill Haus #RPA #Meta Muse #Jamie Dimon
Copyright © DigitalToday. All rights reserved. Unauthorized reproduction and redistribution are prohibited.