Due Diligence With Less Drudgery

Computer Scientists: Ned Gannon, Jacob Mundt and Adam Nguyen.

Sifting through dense legal documents to summarize the main points can seem like mind numbing work meant for a computer. Now it is.

Software developed by eBrevia, a startup based in Manhattan and Stamford, Conn., is able to cut through long-winded documents to extract the key points. Founded in 2011 with technology developed at Columbia, eBrevia is helping law firms and corporate legal departments save time and money and possibly avoid costly mistakes that inevitably arise from human error.

“The tedious work is the bottleneck in the system,” said eBrevia cofounder Adam Nguyen, a 1998 graduate of Columbia College. “The software we’ve developed allows lawyers to focus on practicing law.”

As a junior lawyer who pulled countless all-nighters poring over lengthy contracts, Nguyen saw the benefits of automation first hand. With fellow Harvard Law School graduate Ned Gannon he founded eBrevia and approached Columbia to turn their idea into reality.

Through Columbia’s tech-transfer office, Technology Ventures, he met computer scientist Kathleen McKeown, who heads Columbia’s Data Science Institute and is known for her work training computers to summarize the news. He also met Jacob Mundt, a master’s student in computer science who had previously worked at a startup automating the summarization of health records.

To create the algorithms eBrevia would need to interpret legal records, McKeown and Mundt scanned thousands of contracts and leases made public through the Securities Exchange Commission. After annotating the documents, they slowly trained the computer to recognize and summarize obligations, liabilities and other key provisions buried within. They also taught the computer to recognize variations in American, British and Canadian legalese.

They patented their work and after graduation Mundt joined eBrevia as cofounder and Chief Technology Officer. Last year, the company rolled out software for automating the review of employment contracts, vendor agreements, confidentiality agreements and other types of contracts. In February, they added a new product focusing on real estate leases and will soon turn to other specialized areas like financing contracts tied to mortgage-backed securities and collateralized debt obligations. By spring, they plan to double their staff to 10.

To simplify billing, customers pay to process a set number of documents via eBrevia’s cloud-based software, much like buying minutes in a phone plan. As more companies use the software, the machine learning algorithms are expected to improve.  “The performance of everyone gets better as more people use it,” said Mundt.

Rather than make lawyers obsolete, the software could make them more relevant. “The software allows legal text and data, in a sense, to be visualized,” said McKeown. “That way lawyers can decide what to look at in greater detail.”


550 W. 120th St., Northwest Corner 1401, New York, NY 10027    212-854-5660
©2017 Columbia University