Asset 1
Wall Street’s Big Brother: The AI Software Goldman Sachs And Steve Cohen Are Using To Track Traders

When Tim Estes, 37, created Digital Reasoning as a twenty-one year old with a philosophy degree from the University of Virginia, his startup aimed to use software to bridge mathematics with language and understand human behavior. Sixteen years on, Estes’ language-learning technology is used by a who’s who roster of Wall Street’s most scrutinized firms as they rein in employees and avoid the multi-billion dollar regulatory fines that have plagued the industry since the crisis.

Nashville-based Digital Reasoning, a new member of Forbes’ Fintech 50, was created with the premise that people build recognizable behavioral patterns using language and quantitative technology can be used to define these habits to understand corresponding action. Over time, the firm’s quantitative software teaches itself about patterns thus developing ever more-sophisticated models of individuals’ behavior, which can be used for surveillance ranging from counterintelligence to the combatting of fraud.

Estes’ bet has paid off. A few years after its founding in 2000, Digital Reasoning became an important cog utilized by the Department of Defense during the war in Afghanistan for communications surveillance aimed at gaining intelligence and protecting troops. Digital Reasoning was one of the successful startups that helped the Army comb through a haystack of digital information to spotlight counterterrorism intelligence, or uncover threats to troops on the battlefield.

In 2012, Estes decided to ply this surveillance to Wall Street amid the industry’s litany of billion-dollar fines for actions ranging from insider trading, to client fraud, and market manipulation. In a short period of time, Digital Reasoning’s Army-grade technology has become a staple inside the biggest banks and hedge funds in the world as they use big data and analytics to track employee communications and prevent against the next front-page scandal.

Digital Reasoning uses machine learning technology to scan through millions of emails, instant messages and texts a day on Wall Street to map out ordinary behavioral patterns among the employees of its bank and hege fund clients. This includes screening whom traders ordinarily communicate with, and the types of information they typically send. When behavior is out of the ordinary or falls into a recognizable pattern of wrongdoing, Digital Reasoning then alerts the compliance staff of its clients, who then investigate. Compared with some previous surveillance systems, Estes says Digital Reasoning is far more efficient and lowers false positive rates on alleged wrongdoing by between 95% and 99%.

To read full article, click here.