Agile Analytics Address Ceaseless Change in Financial Crime
February 25th, 2019
“This is the first of a series of blogs authored by Damian Matich, Global Head Fraud Analytics, and Mike Frost, Director of Product Management on analytics. The blog series will address the challenges that financial institutions are facing and how agile analytics step in to lessen the burdens”.
When the Greek philosopher Heraclitus of Ephesus argued in the fifth century BCE for the concept of ‘ceaseless change’ as the fundamental nature of the universe, little did he know that 2,500 years later his dictum would ring true for a new set of disciples, those fraud and compliance professionals battling an aggressive and ever-evolving criminal threat on a global stage.
Criminals continually seek new strategies to launder the proceeds of crime or purloin illicit gains through various types of enterprise fraud. This ever-changing fraud and money laundering landscape demands an agile response delivered within a constrained timeframe to ensure that both the interests of the financial institution and their customers are protected.
Traditionally, financial institutions have sought protection through structured, rules-based detection systems. While this approach enjoys a degree of success, it is usually not capable of meeting the pace in strategy and change demanded by the increasing sophistication of criminal syndicates. Case studies abound where the information exchange enjoyed by fraudsters allowed, as an example, ATM withdrawal limits of banking institutions to be compromised within a matter of hours.
The attraction of analytics lies in its ability to detect fraud more efficiently than rules-based systems and to adapt and learn based on changing fraud trends when appropriate technology is deployed. Furthermore, unsupervised machine learning techniques provide for the detection of anomalous outliers, as has been termed the ‘unknown unknowns’. These are the indicators of new fraud risks and new methods by money launderers to ‘wash’ their illegally gained funds through new channels.
To realise its inherent advantages over rules-based systems, analytics needs data and skilled people to build detection systems based on that data. Data, or more specifically, ‘clean’ validated data, presents a challenge to most financial institutions. An even greater challenge is that posed by the global shortage of people skilled in analytics and data science.
These challenges can be met by providing a secure, external facility for analytic development that incorporates the very latest in modelling tools, working on validated data that is updated and augmented on a regular basis. Such a facility would allow the specialists and subject matter experts within the financial institution a tabula rasa to more effectively meet the detection requirements of their fraud and compliance teams in an agile and timely manner.
This external facility would also pave the way for future efforts on cross-industry cooperation. Criminals do not operate in siloes; they have highly evolved methods of communication and information sharing. Cards compromised in the UK can be printed in Asian factories same day and on the streets of Hong Kong within two to three hours. Financial institutions can only watch with envy at this level of customer service.
In the AML Tech Sprint hosted by the UK Financial Conduct Authority (FCA) in May 2018, this concept of cooperation between industry and partners was highlighted as a necessary element in the fight against organised financial crime. The external facility, probably ‘cloud’ based, would take a daily feed from client monitoring systems and consolidate this data for analysis and reporting. Both client and partner data scientists could then work on the data utilising common tools, focused on building better models either independently or in concert.
It is through this ‘collective intelligence’ that the criminals can be beaten. ‘Better models’ means better detection, which means better outcomes to the fraud and compliance teams fighting the criminals. The external facility would also provide the industry view highlighted by the FCA Tech Sprint, taking as it does data from multiple clients. Ultimately, such an industry view would allow for the provision of proactive advice to participants on fraud and compliance risks in the market.
Most importantly, what the external facility or service provides is an inherent flexibility to provide the opportunity for skilled teams from the financial institution to meet the challenges posed by the ever-changing
modus operandi of the criminal syndicates. By providing the very latest techniques across anomaly detection, supervised and unsupervised learning, network analysis and predictive scoring, what is provided is the adaptability to meet the ‘ceaseless change’ promulgated by Heraclitus over 2,500 years ago.