SST Awards 2018: Best Sell-Side Market Surveillance Product—Nasdaq

It’s easy to see why: Smarts is in use at over a dozen major regulators, including the likes of the Financial Conduct Authority, and at nearly 50 marketplaces, while over 140 market participants make use of the software to safeguard their trading activities. Significant upgrades this year have included the release of its “Lens” module, designed to identify anomalies in trading behavior that may, in some instances, point to wider abusive behavior. It has also continued to expand into new asset classes—most recently, of course, cryptocurrencies, with its adoption by the digital currency exchange Gemini as its surveillance platform of choice.

It has also continued to expand its coverage into more traditional asset classes, through moving into interest-rate swaps, and deepening its presence in the energy and metals sectors. With the go-live of the revised Markets in Financial Instruments Directive (Mifid II) on January 3, too, Smarts has responded by integrating nanosecond-level data models, and further fields as mandated by the regulation, which seek to identify algorithms, market-makers and others.

But where Smarts has really been innovating this year, above and beyond its competition, is in the use of artificial intelligence (AI) and behavioral science. Nasdaq’s 2017 acquisition of Sybenetix gave the Smarts team immediate access to a group of behavioral scientists and analysts, while it has continued to push the boundaries in machine learning. 

The first test case for its work in AI was on its own exchanges in the Nordics, where machine-learning models were deployed to more effectively categorize and prioritize the alerts received by surveillance analysts, drawing on historical patterns to determine whether they were likely to lead to suspicious activity reports and full investigations. The project was successful enough that the technology was later licensed to Hong Kong Exchanges and Clearing in April and is seeking to expand it into its bank client segment through trials with an investment bank.

“The use of machine learning is going to be prolific and it will improve the way we operate across the whole surveillance spectrum,” says Valerie Bannert-Thurner, senior vice president and head of risk and surveillance solutions at Nasdaq. “It will allow us to become significantly more efficient in the way we detect, analyze, investigate, and manage alerts, and it will also allow us to do things differently to before and open up whole new opportunities and approaches to identify suspect individuals and behaviors.”