Machine Learning Saves JPMorgan Chase 360,000 Hours of Legal Work - Technologist
Technologist - The FindLaw Legal Technology Blog

Machine Learning Saves JPMorgan Chase 360,000 Hours of Legal Work

Here's a sobering fact for lawyers: Last year, a machine learning program used by JPMorgan Chase saved the company 360,000 hours of work, work that would normally be performed by lawyers and loan officers.

That's more than 41 years worth of nonstop legal work -- all handled in a few seconds, by some well-designed software.

"A New Era of Automation"

The lawyer-eliminating program, called COIN for "Contract Intelligence" automates the interpretation of commercial loan agreements, a "mind-numbing job" that would normally be handled by attorneys and loan officers, according to Bloomberg News.

"The software reviews documents in seconds, is less error-prone and never asks for vacation," Bloomberg writes.

The program relies on machine learning algorithms, learning "by ingesting data to identify patterns and relationships." It's part of what Bloomberg describes as "a new era of automation," one that's "now in overdrive as cheap computing power converges with fears of losing customers to startups."

JPMorgan is already looking for more ways to use the technology. The bank plans to use machine learning to help deal with complex legal filings, for example. It's also considering utilizing the technology to help it interpret regulations and analyze corporate communications.

JPMorgan Invests Heavily in Tech

Of course, staying on the cutting edge of technology isn't cheap. JPMorgan employs 20,000 developers and has a technology budget just shy of $10 billion. The bank's total technology budget takes up 9 percent of its projected revenue, Bloomberg reports, twice the industry average.

The company is deploying its technology on a variety of fronts. In addition to automating the review of commercial loan agreements, it's investing in blockchain technology, increased cybersecurity, using email analysis for sales leads, and even creating simple bots. Bot programs will allow the company to automate routine tasks like resetting employee passwords. Already, bots are expected to handle 1.7 million requests at the company this year -- a job that would have otherwise required flesh-based employees.

JPMorgan isn't worried that its focus on technology and automation will end up making it a "post-human" company, though. "People always talk about this stuff as displacement," says Dana Deasy, JPMorgan's chief information officer. "I talk about it as freeing people to work on higher-value things, which is why it's such a terrific opportunity for the firm."

Related Resources: