SYNAPSE – A SOLUTION TO FINANCIAL CRIME

Introduction

Alpha Reply exists to help Financial Institutions solve complex business problems using data, models, and technology.

It needs the right platform to develop innovative ideas that we could deploy across a large financial services customer base.

Reply offered the right platform based on its three characteristics:

• A DNA deeply rooted in tech
• A proven culture of innovation
• An appetite to take risks in order to invest in innovative ideas

Our strategy from the start has thus been to:

• Help our clients solve immediate business, regulatory and technological challenges they are facing.
• Invest in building technological tools that solve widespread existing problems facing our client base.

Fast forward I am now starting a series of blog-posts discussing a software solution we are currently developing that allows banks, e-money institutions, and payment processors to test and stress test their financial crime controls.

The Problem: Lack of Current Controls

Let me start by defining financial crime and its controls: Financial crime refers to any crime involving an individual or group of individuals using financial products offered by banks to:

• Commit fraud (essentially stealing money from the bank's clients)
• Engage in money laundering, financing terrorism or facilitating human trafficking.

Financial crime controls are the set of processes banks put in place to prevent – or mitigate the impact of – financial crime. It is quite complex to test the effectiveness of the control for two quite obvious reasons:

• Financial criminals tend to be creative and have become very good at continuously finding new ways of exploiting the bank's products to commit crimes.
• In most cases, where there is no actual financial loss incurred by another party, banks have virtually no way of assessing how effective their controls are (i.e. a money launderer having successfully laundered money through a bank will most likely not go back to the bank to brag about it…).

Banks and regulators have launched numerous initiatives to mitigate the above, such as forums and bodies (FATF, …) who facilitate the communication of best practices and the dissemination of existing and emerging financial crime schemes (known as typologies).

While these initiatives have largely improved these controls, financial crime is nevertheless still prominent. Its consequences are hugely detrimental to society and banks themselves where regulators have imposed massive fines to financial institutions with insufficient controls.

Our Proposal

It is in this context that we have developed a software solution, Synapse that simulates financial transactions by combining:

• Good agents, engaging in non-financial crime-related activities
• Bad agents, engaging in financial crime-related activities implementing many known typologies.

Our goal is to make this agent-based simulation tool a key part of how banks go about testing the effectiveness of their controls, leveraging the fact that they can directly test for transactions that:

• They should have flagged but haven't (false negatives)
• They shouldn't have flagged but have (false positives)

It is also a powerful tool for rapidly implementing emerging typologies (using our behavior scripting engine) escalated by our users and documented by various bodies to directly test for vulnerabilities in control environments. As with any complex problem, the development of our solution is a continuous work-in-progress.

In the coming months, I will be posting weekly content about our tool, discussing various aspects, covering technological, methodological, business, and infrastructure aspects of it.

If you would like to partner with Reply, please contact us at .