JPMorgan rolls out robots to scrutinize banker travel, expenses

JPMORGAN CHASE & CO. has begun using machine-learning technology to process expense reports and determine whether they comply with company policies. / BLOOMBERG FILE PHOTO/PETER FOLEY
JPMORGAN CHASE & CO. has begun using machine-learning technology to process expense reports and determine whether they comply with company policies. / BLOOMBERG FILE PHOTO/PETER FOLEY

NEW YORK – JPMorgan Chase & Co. deal-makers on the road have another reason to resist the mini bar: The robots are watching.

The bank has started using machine-learning technology to process expense reports and determine whether they comply with company policies, according to Lori Beer, JPMorgan’s chief information officer.

“We basically have eliminated manager approvals,” she said Wednesday at a conference in New York. “We’re doing 100% of audit through a machine-learning model that makes sure that as we process travel and expense reports, they’re in alignment with our policies.”

Machine-learning is a type of artificial intelligence that uses data analysis to spot patterns and improve itself over time, making better decisions without being explicitly programmed.

- Advertisement -

While almost every industry wrestles with policing expenses, it’s been a particular bane in finance, where employees with refined tastes spend much of the year on the road and meeting clients. Wells Fargo & Co. last year fired or suspended more than a dozen employees for allegedly falsifying expense reports, The Wall Street Journal reported at the time.

Yet putting pressure on managers to spend more time scrutinizing reports, or hiring auditors to do it for them, basically adds to expenses. With artificial intelligence, JPMorgan is “taking some bureaucracy out of our managers’ hands,” Beer said.

The strategy is just one of the ways the bank has been trying to harness cutting-edge machines to become more efficient, reduce fraud and improve experiences for employees and customers.

The bank last year hired Manuela Veloso, Carnegie Mellon’s head of machine-learning, to help the bank build on existing work. In April, CEO Jamie Dimon told shareholders in a letter that machine-learning could help the bank save $150 million by better detecting credit card fraud.

Michelle F. Davis is a reporter for Bloomberg News.

No posts to display