Check out the article to learn how data plays a part in preventing fraud in a FinTec Buzz exclusive interview.
Please tell us about ComplyAdvantage and the types of organizations you work with.
Over the course of our history, ComplyAdvantage has established a reputation as a trusted partner to anti-money laundering (AML)-regulated companies, with services that address the unique needs of financial institutions, fintechs and payment companies. We’ve grown beyond simply AML use-cases to provide AI-driven financial crime risk data and AI-based fraud detection technology to institutions around the world — giving regulated businesses the information and insights they need to detect and prevent money laundering and other financial crimes. With solutions including Know Your Business, Fraud Detection and more, we enable more than 1,000 enterprises in 75 countries to rapidly understand and evaluate the risk of who they’re doing business with and create a safer, stronger financial system. As we continue to grow and expand, our comprehensive approach will help an even broader range of customers, from promising startups and dynamic scaleups to large regional and international enterprises.
Can you tell us about your role and background?
I’ve been at ComplyAdvantage for about six months and have been focused on innovating and improving the data supply chain that is the foundation of so many of our core products. This includes massively increasing the scope of data typology that we can consume, building smarter ways of extracting relevant facts from that data, and improving the way we connect those facts together to surface patterns of financial risk for our clients.
Prior to my role at ComplyAdvantage, I was vice president of data at GoCardless, where I led the creation of the firm’s data capabilities and organization that underpin its payment intelligence products. I’ve also worked at consumer-focused tech companies where data was a major focus.
You’re the first chief data officer at ComplyAdvantage. Why is this type of position important? What types of companies should consider adding a chief data officer to their executive team?
Institutions today need to know that the people or organizations they choose to do business with are legitimate and that the transactions they facilitate are legal. Data is a key part of that, with up-to-the-minute information at the core of the fight against money laundering and other forms of financial crime. Because our clients rely on ComplyAdvantage for comprehensive, valuable and actionable data, establishing a “chief data officer” (CDO) position is a natural and important fit for our business: reflective of the centrality of quality and comprehensive insights to our mission and products.
As for other companies considering a CDO and expanding data-related positions: If you are going to build excellent data-driven products, you need to be able to build teams of engineers, data scientists and product people who work closely together to bring the insights from all three lenses to your product development efforts. The role of data scientists is to find signals amid the noise, while engineers focus on how to build a reliable machine, and product people work out what is most valuable to our customers. The three of those roles have to work in concert to get to the best solutions that address customers’ problems. A CDO works alongside the chief technology officer and chief product officer to achieve this, plus works across the rest of the organization to ensure that data-driven decision making happens effectively in non-product areas as well.
We live in a sea of data that’s growing exponentially each day. How do you make sure you’re sourcing the best, most up-to-date data for ComplyAdvantage’s clients?
It all comes down to a combination of people and technology. We have domain experts who understand how to source the best data and technicians who know how to take that data and scale it. The most important issue is trust — our clients need to know that they can trust the data that is at the foundation of our solutions and that they are using to manage their risk. We use a combination of human expertise and automated techniques to ensure that the sources of our data are trustworthy and accurate.
With so much data available to financial services companies today, how can AI help them manage the data deluge and draw actionable insights? (And while AI is very important, why is so much conversation today focused around AI and not the underlying data?)
If you zoom out, often AI is being used to make decisions that in the past you would have relied on humans to make. The difference is that you are likely making decisions at a rate and scale that would be impossible to do with people. At the same time, you need to ensure the outputs are accurate, that you don’t unintentionally introduce bias into your decisioning, and that the decisions made by the AI model are trustworthy. If an organization is going to partner with a provider to integrate AI into its business, it is important to know that they have built responsibly. Institutions need to hold their suppliers to account to make sure that the decisions that their technology is making are auditable, testable and explainable.
This is why there needs to be more of an emphasis today on the data that is being used to train the models. A model will never perform better than the underlying datasets, which is why investment in data governance is just as important as investment in the technology.
What is the role of data in fighting fraud? How does ComplyAdvantage help clients with this?
At the root of the fight against fraud is the ability to spot anomalies and patterns of data — and to do that quickly and accurately enough that fraud can be identified and prevented. Solutions need to take into account multiple sources of evidence from transactional data (the spending patterns that take place between criminal and victim) to behavioural data that can be indicative of a crime that is about to happen (e.g., password or address changes) to other available risk indicators that can be derived from adverse media or other publically or commercially available data sources.
Fraud is an ongoing fight, and data and technology need to address this always-on struggle by constantly adapting to new threats and to the criminals who are constantly testing new technologies to find and exploit weaknesses in the system. You are never done. You can never create a data model to address a particular fraud typology and say “Well, that problem is solved.” You need to always be iterating and evolving to address new patterns of behavior.
What is the role of data in conducting customer due diligence?
Businesses, whether they are regulated or not, want to manage the risks that they are exposed to. While transactional data is one part of this, risk management also encompasses knowing things that have happened outside of the transactional space. An example of this is adverse media data, which provides color and depth by showing the customer’s connections to associates or events that could be indicative of criminal behavior.
Can you tell us about the types of data and intelligence ComplyAdvantage provides to clients?
In addition to the adverse media data mentioned above, we also provide the most accurate and timely information on national and international sanction watchlists, as well as an up-to-date list of politically exposed persons (PEPs). With this landscape changing on an almost daily basis over the last 18 months, the importance of timely updates for AML-regulated businesses such as financial institutions has never been more critical. In addition to the proprietary data that we own and manage, we can incorporate our clients’ data as well — providing them with a broader view of the financial landscape and helping them manage their risk exposure while also protecting their own customers from criminals.
What makes a good data scientist? What kinds of qualities and analytical thinking do you look for in your team?
The best data scientists are constantly questioning and iterating. The nature of our environment means that it is always expanding with updated information as well as new sources of data. It’s our job to first question that data’s accuracy and, second, to explore if it can be used to create better results in existing models or, indeed, if new models are needed.
In addition to seeking candidates who possess the technical aspects of the data scientist role, we also look for people who have interest in, and an understanding of, the domain we work in; have strong ethical principles; and are able to navigate the increasingly complex environment, as AI becomes more performant and more pervasive over time.
What do you think the finance industry should be focused on right now?
Just as AI is providing new opportunities to identify and prevent financial crime, it is also giving criminals new ways to separate people and businesses from their money. All industries — not just the financial sector — should be concerned about the threats of deep fakes and synthetic identity fraud. Education campaigns; stringent identity verification processes; accurate, comprehensive, real-time data; and trusted technology partners can help manage and mitigate always-on risks such as these, so organizations are resilient and future-ready.
Stay Ahead of the Financial Curve with Our Latest Fintech News Updates!
