Within the broad universe of information that constitutes the data generated every day, both by financial institutions, and authorities, there exists an extensive set of information relevant to understanding normal and abnormal transactional activity. Having access to this data and understanding this distinction is key to the development of any technology capable of accurately detecting patterns of financial crime.
For a while, the industry has been asking for more information from the public sector to support efforts to prevent and detect financial crime, but information sharing is typically confined to round-tables or in the course of case-by-case exchanges. On the whole, the industry still lacks systematic access to the broader, labelled dataset (i.e. data on true positives) produced by authorities that would enable the industry to effectively reverse engineer patterns of illicit behavior. Information sharing and active collaboration across all concerned parties within the financial system is critical to successfully combating financial crime.
Absent the provision of quality labelled data, banks are left to use overly simplistic detection mechanisms, often spending vast amounts on managing false positives and only identifying suspicious activity through the vigilance of their staff rather than through their systems. Such a situation is suboptimal for various reasons, ranging from the fact that not all staff are equally trained and capable of detecting suspicious activity, to the fact that manual escalations tend to focus on retail and commercial banking, leaving institutions exposed in other areas, to the fact that a large dataset which could help solve this problem is being under-utilised.
This situation is not solely due to deficiencies in current technological capabilities.
The gap between current control standards, and what is actually possible exists in part due to the industry’s lack of access to labelled data on financial crime outcomes.
Were this labelled data to be provided and processed using state-of-the-art data analytics and machine learning technologies, otherwise invisible trends and patterns indicative of criminal activity would be made identifiable. When contextualised alongside transactional chains leading up to, and subsequent to a noted suspicious transaction, data on actual investigatory outcomes would provide effective confirmation of a given pattern. Across billions of financial transactions, this would allow for the detection of patterns known to be indicative of illicit activity, and the identification of patterns that were previously unknown. As an iterative process, the strength of this identification capability would only grow over time.
This data would provide an unparalleled view on financial crime risk exposure at a systemic level, adding context to future transactions, and in doing so, enhance the analytical capabilities of the private sector and authorities alike.
Building a control framework such as the one discussed above is not out of reach. All the necessary tools and capabilities are directly in front of us. All that is required is the will power to construct it, and a commitment to work in concert towards the common goal of more effective and more modern financial crime prevention. If ridding the world of financial crime is the goal, cooperation and an appetite for change are key. Elucidate has invested in building the required technology and dataset, and has continuously advocated for a more concerted effort to increase transparency so as to more effectively combat financial crime.