By Mallinath Sengupta, Chief Executive Officer, NextAngles, a provider of artificial intelligence-based systems for compliance applications
For banking and financial institution executives – and for their investors – 2016 has begun on a sour note. From the largest money center banks to small local institutions, double-digit earnings declines were commonplace in the first quarter, as banks suffered from slower loan growth, challenging capital markets, and elevated provisions for energy credits, among other factors. For the full year, some analysts are forecasting a 20% drop in average in earnings for the top U.S. banks.[1]
Among the forces burdening banks’ financial results is the rising cost of regulatory compliance. The added expense, for everything from anti-money laundering (AML), to know your customer (KYC) requirements, to the compilation of voluminous data for stress tests, has been estimated at up to $4 billion per year at some big banks.[2] And that does not even include fines for compliance breaches.
While banks cannot control many of the factors dragging down earnings, compliance cost is one area within management’s control. In particular, advanced artificial intelligence (AI) can be applied to key compliance processes, providing smart, technology-enabled solutions. At NextAngles, we believe that AI has the potential to be more precise, at a lower cost, than the traditional approach of “throwing more bodies” at compliance needs.
To understand the real benefits of artificial intelligence, it is instructive to look at why it is so difficult and costly to manage the compliance burden in the traditional way. In the area of AML, for example, financial institutions have transaction monitoring systems that generate alerts when potentially unusual activity is detected. In order to be thorough and avoid heavy fines, the systems are extremely sensitive and thus generate large numbers of false positives. This means that compliance staff must scrutinize each alert, investigate the activity, and determine whether it is unusual and rises to the level of being reportable in the form of a Suspicious Activity Report (SAR). The problem is magnified since financial institutions typically have multiple systems, making it necessary to compile data from numerous sources in an investigation.
AI can solve this problem by creating domain-centric models that replicate the “real world” of banking and regulatory compliance. The advantage of AI systems is that they are able to perform tasks that normally require human intelligence, such as pattern recognition and even lower-level decision-making.
Importantly, AI enables the creation of “learning systems” that can become more expert with each subsequent investigation. AI does not replace human intelligence, but it can perform lower-level knowledge functions efficiently, enabling team members to save their time and effort for higher-level decision-making.
While this example has focused on AML compliance, AI systems have applications for other compliance areas, such as KYC, insider trading monitoring and Basel III liquidity solutions. Based on pilot projects, we believe the use of AI in compliance applications can produce up to a [30%] reduction in costs and yield three times faster throughput.
It is still early days in bank’s use of AI for compliance functions. One possible reason is that the current focus of their digital technology spending lies elsewhere, in areas such as enabling mobile banking, or automating some loan underwriting functions. It may be easier to get budget dollars approved for revenue-generating activities, rather than “overhead”.
AI-based compliance systems have the potential to pay big dividends for banks and financial institutions, however. Not only can AI help reduce the risk of costly fines, but it can also stem the growth of compliance costs and permit the savings to be redeployed toward revenue-generating investments.
We also believe that next-generation artificial intelligence systems will help drive top- and bottom-line growth. The same capabilities that today allow AI to detect patterns of possible fraud should enable tomorrow’s systems to detect patterns in customer needs and allow banks to target sales efforts more precisely. Until then, however, AI provides a smart platform for building faster, more accurate and more cost-effective compliance solutions.
References
[1] Reuters, April 11, 2016
http://www.reuters.com/article/us-banking-results-idUSKCN0X70VU
[2] Financial Times, May 28, 2015
http://www.ft.com/cms/s/0/e1323e18-0478-11e5-95ad-00144feabdc0.html#axzz478sI8EzJ