Cognizance of the Looming Threat
A recent disclosure by US law enforcement authorities reported about three Colombian nationals who used Hong Kong banks to launder at least US$5 billion from a global narcotics business . That’s more than the GDP of a few nations such as Sierra Leone or Barbados. In this instance, the money was used to largely buy counterfeit products from China and Hong Kong, and ship them to the rest of the world. But it could just as well have been used to pay for weapons or fund terrorist activities.
Money laundering is posing one of the biggest threats to societies and economies across the world. Banks are naturally becoming the first line of defence against the dangers of financial crime. Increasingly, regulators are asking banks to gather information about their customers and their transactions that will help surface, identify and prevent potential illegal activities. This has set off an intense chase for data and supporting documentation, inconveniencing customers, causing business delays and further adding to operational costs. The Know Your Customer (KYC) process mandated by regulators is turning into a double-edged sword: while augmenting anti-money laundering (AML) processes, financial institutions are becoming more vulnerable with increasing instances of customer privacy breach and data theft.
Uncovering the Challenges
Global banks and financial institutions need to comply with the ever increasing regulations, protect their customer data, implement AML technologies and processes, ensure they are not exposed to reputational damage and keep the cost of doing all this under control. While there is no common estimate of what these safety measures cost the financial industry owing to various qualitative factors, they do result in increased lending rates and service fees.
Current compliance processes are intensely manual in nature. This means collating information from diverse and disparate sources to include origin of customer, nature of business, holding structure, political affilliations and connections, financial history, referrals, sanctions, indictments, liabilities, news reports, teller/automated clearing, house/wire transaction records, frequency and volume of transactions, currencies and geographies across which transactions take place and so on. The task encompasses country to country, bank to bank and people to people transactions. Almost 80% of this task is done manually by analysts, often spending days and weeks in acquiring, searching, documenting and reporting precise information about people and organizations - their financial standing and their reputations rather than in actually analyzing to make informed decisions. Finally, when the analysis is done, it may have been weeks, months or even years after a money laundering event
Need for AML Solutions 2.0
Current AML solutions have evolved their ability to manage structured records and data. What the industry needs is an automated way of searching and sifting through the growing volumes of unstructured data along with the structured data. The solution must then have an additional layer of analytics that understands and unravels monetary trails using sophisticated models to uncover relationships and predict patterns of suspicious activity. The solution calls for a combination of technological expertise and domain knowledge.
Simply put, the system must look for appropriate and high quality data independently across defined sources (public, syndicated and third party), interpret unstructured data (such as descriptions and narratives, opinions and notes), synthesize and standardize it, create profiles that are aligned with regulatory requirements, use a set of industry-and-business-specific models to query and analyze the corpus of data and immediately flag activities that analysts need to take decisions on.
Making Smart Solutions Smarter
The problem is that even the most sophisticated systems deployed across financial institutions are prone to making errors when dealing with unstructured data. For example, systems may get confused between social identity, tax registration, phone and even travel document numbers. Cognitive Process Automation (CPA) and Artificial Intelligence (AI) engines developed under the guidance of industry compliance and risk practitioners offers a reliable, fast and scalable solution to this problem.
CPA is driven by a set of instructions describing processes. CPA-based systems can literally “learn” these instructions by observing the actions of financial analyst and by examining a historical database of actions and behavior. As the system processes more data, it learns continuously to develop new rules and associations, presenting them to analysts for approval and inclusion in the rules engines that analyze data.
There is no doubt that growing regulatory requirements, the massive amounts of structured and unstructured data and the cost of managing these present a challenge to the industry at large. The risk associated with poor AML practices cannot be underestimated simply because crime methodologies are getting highly sophisticated and need equally intelligent counter measures.
1. http://www.scmp.com/news/hong-kong/law-crime/article/1857155/ laundering-ring-pumped-billions-drug-money-through-hong
2. As one way to view the broader impact, research by the International Monetary Fund has suggested that proposed reforms could reduce economic output in advanced economies alone by ~3% during 2011–15: https://www.imf.org/external/pubs/ft/wp/2012/wp12233.pdf