OpRisk Awards 2016
As volumes of transaction and client data grow seemingly exponentially, so too do the challenges that governance and risk compliance teams face in making sense of it all, and in detecting any fraud or money laundering that may be going on.
"For a compliance application, it's all about understanding the data," says Don Ryu, director of product management at Oracle Financial Services Analytical Applications. "You can have the best visualisation, the best analytic modelling tools, but if you don't have control over the data, they're no use."
At the foundation of Oracle's offerings across all 46 of its financial services analytical applications – including its Financial Crime and Compliance Management product – is a common analytical data model. As Ryu explains, the California-based technology firm takes a bottom-up approach to data modelling, ensuring the data is well structured and standardised, allowing it to be shared almost seamlessly between clients and applications.
You can take data from client A, and take it to client B's site, run the analytics and get 99% of the same results. That's a very unique feature that only our product offers
Don Ryu, Oracle Financial Services Analytical Applications
"The data model is pretty much set across clients, so the usage of data is consistent. We take a ground-up approach: you can take data from client A, and take it to client B's site, run the analytics and get 99% of the same results. That's a very unique feature that only our product offers," says Ryu.
A particular challenge financial institutions face in dealing with increasing volumes of data is the growing number of ‘false positive' alarms produced by many risk-monitoring systems.
"A big theme for our clients is around improving the productivity of the compliance function," says Ryu. This means ensuring that applications provide accurate, timely information, with context that allows the compliance team to identify suspicious activity.
The Oracle Financial Services Behavior Detection Platform uses link analysis algorithms that are able to build networks of entities based on common attributes, such as addresses and tax identifiers, and types of transaction activity, which can give users visibility into networks of which they previously may have been unaware.
"By definition, money laundering cannot be identified by a single transaction," Ryu says. "It's a group of activities with a specific intent to hide the source of funds. Our application has strong correlation engine capability to bring relevant cases together to provide the context needed."
Non-traditional data sources
Oracle is also investing in the use of non-traditional data sources, such as negative news flow, emails and social media information to add context – both to aid any internal observation, and to provide to regulatory authorities.
"Clients also want to see overall robustness: how much can you trust the system and ensure it provides full data lineage that you can provide to regulators, or to internal or external audit," says Ryu.
As he notes, financial crime protection doesn't stop with detection. An important feature of Oracle's crime and compliance offering is the interface and workflow that allows for efficient investigation and resolution of alerts and cases.
"Alert management, case management, all the way through to regulatory reports are all managed in a single place, and are transferred automatically, reducing the possibility of manual error," he says. "It's all about delivering the right productivity to users."
The week on Risk.net, December 9–15 2017Receive this by email