Michael Shashoua: Seeking Smarter Applications

Two trends that Inside Reference Data covered recently—the rise of “regtech” (regulatory technology) as a spin-off of the development of fintech start-ups, and the increasing application of artificial intelligence and machine learning to data management—are both about how smarter application of technological capabilities can improve the industry’s grasp on data.

In the UK, regtech sprung out of fintech because half of new start-ups fold within five years, and fintech ventures are largely start-up in nature. The survivors are those that focus on back-office, compliance and post-trade functions. Compliance demands make regtech services for back offices a surer thing as a start-up venture.

Users of regtech services want to get out ahead of compliance requirements, and don’t want regulators’ demands driving what regtech developers offer. As it stands, however, the regulators may have the upper hand in defining what regtech developers do offer.

Regulators’ push for more frequent or even real-time access to up-to-date data is really what’s “ensuring that the next generation of solutions comes through,” says Ian Manocha, CEO of UK-based Gresham Computing.

Regulators’ demands could be enough to fuel regtech successes. Regtech end-users or potential end-users who want a say in how regtech works and develops should engage with its creators now. Whether the regulators or the regulated end up with a tighter grasp on how real-time compliance data is handled, through newer regtech offerings, is the key question.

In data management, machine learning can be innovative if it proves to be a new way to gain previously impossible insights.

Machine Learning Expands

The industry appears to be further along with deployment of machine learning efforts for financial data management. In March, Bloomberg launched its Liquidity Assessment tool and crowdsourcing investment platform StockViews secured funding to apply machine learning and artificial intelligence to enhance its crowdsourced research on companies. In February, performance measurement and analytics vendor Velocimetrics announced the addition of machine-learning techniques to its market data quality solution.

These are just the newest entrants. WorkFusion has for some time been applying artificial intelligence to automating repetitive data processing tasks. IIROC, Canada’s major self-regulatory organization, completed a machine-learning project to segment market participants, showing that regulators also can make use of other new technological capabilities aside from regtech.

And several other companies, including AltX, Dataminr and Verafin, have been using machine learning to yield greater insights from data, whether for portfolio managers or for compliance purposes.

These add up to quite a few machine learning ventures, and could be just the tip of the iceberg. The machine-learning field boasts more providers than regtech at this stage, but still has unanswered questions about whether its efforts use artificially intelligent processes effectively to derive insights from data, and therefore produce actionable intelligence.

Using artificial intelligence to arrive at new insights is a more complex and challenging task than merely getting better quality results through the automation of data processing.

At ISITC’s annual conference last month, Jeff Zoller, chair of the financial industry operations group, broke down the challenges for new technology in the industry into three parts: iteration (articulating the problem that needs to be addressed); innovation (producing a new capability or way to address that problem); and disruption (completely changing the established processes that gave rise to the problem). Few efforts truly qualify as disruptive, Zoller said. 

In data management, machine learning can be innovative if it proves to be a new way to gain previously impossible insights. It’s not yet evident whether those insights are enough to make machine learning disruptive. Regtech, on the other hand, could go far just by correctly establishing what needs to be done.