The Game is Afoot: The Simple, the Complex, and the Deep Learning Solutions for Compliance Programs
Current compliance methods relying on centralized rules-based transaction monitoring are plagued with false-positive results. Suspicious Activity Reports (SARs) are generated from a vast minority of compliance cases, and the end-user (law enforcement) rate-of-use is somewhere between unknown and abysmal.
Remaining compliant, especially with transaction monitoring regulation, is a beast unto itself with its own metrics for success, its own financial burdens and (potentially catastrophic) risks of fines, penalties, and reputation loss should a program be out of compliance.
However, identifying criminal networks requires a change in approach that goes even further and transcends the boundaries of compliance-focused frameworks - as these do not necessarily pay sufficient attention to non-obvious risks or threats that do not or cannot be identified with current rules and thresholds.
“We’re not here to find the bad guy, Brandon. We’re here to be in compliance.” I repeatedly encountered this position throughout the compliance industry (read as Anti-money Laundering, or AML, Fraud, and Cyber organizations) during my time in corporate banking. I have never quite understood why one cannot focus on these simultaneously. At FNA, we are on a mission to provide solutions that enable our clients to approach both of these objectives as equal priorities.
- Simplified Investigative & Multi-Source Knowledge Graphs -
Starting simply, FNA allows KYC based investigations (such as, negative news and sanctions screening) and routine transaction analysis to take place visually so that links between high risk and low-risk entities can be intuitively identified through visualization. FNA enables multi-source data investigations displayed as a combined graph. This ensures that assessments are consistent across business units and based on holistic data.
Figure 1 (Below). Bank Saderat Iran query results if only using a single source, in this case the Office of Foreign Asset Control (OFAC) Specially Designated Nationals (Sanctions) list. Note the lack of linkage between two sanctioned entities for the same firm.
Figure 2 (Below) shows the same firm overlaid with Legal Entity Identifier data (merged and visualized in the FNA Platform) about subsidiary ownership and reveals two additional branches worldwide owned by the sanctioned entity, but not present in the SDN list. Therefore a sanctions screening team could easily miss these relationships.
Additionally, our knowledge graph solutions for analyzing merged datasets with graph algorithms enable us to identify previously unknown links between entities of interest or concern and measure influence, diffusion, distance, or other relevant facets in one visualization.
Figure 3 (Below) shows the linkages between Cambridge Analytica and current politically exposed persons using open source intelligence web scraping and structured business ownership reporting from Companies House and Bloomberg.
This allows for visual KYC, CDD, and PEP/Sanctions investigations using multiple data sources internal and external to a firm and purpose-built query capability.
As seen above, tracing ownership and business relationship relationships stemming from Cambridge Analytica (pink node in the center of the graph) can quickly lead to interesting relationships and affiliations that have PEP and other political considerations. Note the different sizes, colors, and opacities of the nodes and links, governed by out-of-the-box graph analytics that automatically create a visualization showing the strongest linkages and most influential entities on the graph.
In the next part of this series, we will focus on how to augment centralized monitoring solutions with an aim to increase monitoring effectiveness and efficiency.
Interested in how FNA can assist with your investigations or information management requirements? We are on a mission to support companies with these needs in the challenging business climate. Reach us at firstname.lastname@example.org if you would like to learn more.
By Brandon Smith, April 9, 2020, Published on FNA