Intelligent Machines Make Better Compliance Decisions

A UK CEO arrives at work at dawn, eager to close a deal with an Asia-based trading counterpart. In an hour, she will be joining many other bidders eager to snatch the same deal. Comfortable in her office chair now, she punches her code to access the company’s online accounts, only to find a message pop out on her screen saying all company accounts are frozen.

In a state of shock, she realizes the nightmare she had dreaded most has just happened. Her ongoing negotiations with the Bank regarding the company’s compliance issues and related transactions failed to produce the resolution she hoped for.

Because now she couldn’t deliver the confirmation contract bond to the Asian seller, this golden trading opportunity has slipped away. Calling the bank to remedy the situation would also be useless because the National Crime Agency (NCA) has ordered the accounts frozen. The CEO then realizes that her company will soon go under a lengthy and costly money laundering and terrorist financing investigation, which will severely limit her ability to conduct business, and if the investigators prove a case of compliance negligence, a hefty fine could also be slapped on the company.

Her bank’s Anti-Money Laundering (AML) system had been generating alerts based on specific compliance algorithms before the case was elevated to the NCA by the bank. These compliance systems are optimized to detect irregular and suspicious transactions based on a diverse set of transaction scenarios. They first produce suspicious alerts that the Bank would only elevate to suspicious activity reports (SARs) if the company cannot supply proper explanations.

The SARs are then sent to a regulatory crime unit for assessment and action. The crime unit then decides whether to freeze a company’s accounts or not. If the bank ignores the suspicious reports, it risks exposing itself to massive regulatory fines and negative media coverage.

Each bank installs different financial compliance algorithms, developing those that serve their unique jurisdiction best. In this CEO’s case, the Bank’s compliance system must have been triggering suspicious transaction alerts that her company couldn’t explain adequately.

Regulatory fines on banks have been rising fast and often run into billions of dollars. There’s also the threat of jail time for bankers who ignore regulatory warnings. Because most banks want to avoid dealing with the heavy hand of regulators, they forward a lot of suspicious activity reports to the financial crime unit of their jurisdiction to vet them out. Passing SARs to the regulator also reduces a bank’s workload and noncompliance risks. Within 18 months, for example, the UK’s Financial Intelligence Unit (UKFIU) had received a record 630,000 SARs from banks in its jurisdiction.

The Promise of AI

Powerful new technologies now offer renewed hope to both bank and regulator for more effective and efficient compliance management. These new technologies address the challenges of accurately identifying suspicious client behavior. Through the application of artificial intelligence (AI) and other related technologies, the flood of bank transactions can now be assessed and vetted in real time. These systems can autonomously analyze millions of transactions, categorizing them into groups based on a diverse set of risk values.

In another scenario for our CEO, her bank’s compliance system could have cross-analyzed data from different sources before rushing SARs out to the regulator. Using multiple sources of data and intelligent analysis would interpret the Company’s transactions differently and as such would not have generated the SARs in question.

Relying on legacy compliance systems means officers at the bank will have to go through each suspicious case manually, which is near impossible. For most banks, such routines are complicated, costly and time-consuming. This is where AI-systems come to play, intelligently cross-analyzing the massive data feeds produced by transactions; assigning them risk values and effectively reducing compliance workloads and failures.

These intelligent solutions vary and come semi or fully autonomous. They develop analytical models based on analyzing the history of an entity’s behavioral patterns. Often, these AI-powered compliance systems can detect suspicious transactions that most compliance officers wouldn’t see.

The promise of applying advanced artificial intelligence (AI), blockchain and the Internet of Things (IoT) systems is ameliorating the problem of runaway SARs and false positives. There’s no limit really to how much data a bank can acquire and analyze to support and validate transaction events.

Through IoT, compliance systems can access data related to manufacturing, transportation, insurance and more, deriving a multi-dimensional insight into a customer’s financial behavior in real time. Such an approach to anti-money laundering can significantly mitigate a bank’s risks and effectively lower the cost of compliance.

In the very near future and as AI systems take over in most banks, the debacle our CEO had to experience would become a thing of the past.

Belal Hejawi                                                                                              

Content Development EastNets