Jump to content
  • Intelligent Fraud Management


    All banks experience fraudulent activities across their retail products and services. For the banks this focuses on a couple of main areas i.e., Anti Money Laundering or AML.  and Sanction Compliance

    Introduction

    All banks experience fraudulent activities across their retail products and services. Financial crime in retail banking is mainly targeted in two areas: Payment Services and Card activities. Most people are aware of card fraud that is associated with the theft of physical cards, or the theft of card details that enable fraudulent purchases and cash withdrawals. Payment services fraud is less visible to consumers but is very visible to banking regulators because of its links to organized crime. For the banks this focuses on a couple of main areas:

    • Anti Money Laundering or AML. 
    • Sanction Compliance. Countries can be sanctioned for a number of reasons and different types of sanctions are enacted both globally and by specific jurisdictions.  Sanctions can be specific to an individual or applied to an entire country and commonly include a limitation or exclusion of payment services.

    But why do banks invest so much money into this? This is not just because of the direct losses, it's also because of the huge fines associated with noncompliance and the loss of reputation for the bank.

    Financial crime prevention and investigation are very costly in themselves. It takes large teams of investigators to check the huge volumes of suspicious transactions that most systems detect. Typically after assessing the suspect transactions the vast majority turn out not to be fraudulent, as the banking systems have created a false alert from the rules that were applied. In banking terms this is considered to be a "False Positive", the reduction of the false positive rate is the "Holy Grail" of fraud detection.

    Three types of criminal activities in retail banking

    Card Fraud

    Card Fraud is usually a potential loss that ultimately stays with the credit card division or affiliate of the bank. As the card payment business is a three or four-party system [The cardholder, the card issuer, the card acquirer, and the merchant], preventing fraud is a challenge. The most typical fraud patterns are:

    • Chargebacks: A person refuses a transaction as they claim to not have authorized the transaction. Where the merchant has followed the card policies and the cardholder can make a plausible claim that they were, for example, not present in the shop, the result is a loss for the bank.
    • Merchant Identity Fraud is common online. This is where a customer believes they are purchasing from a well-known online supplier but in reality, the web shop is being run by somebody else.
    • Buyer Identity Fraud is where someone makes a purchase with the identity of somebody else. 

    Payment Service Fraud

    Payment Service Fraud has been around since the advent of banking. Fifty years ago the first US regulations were introduced to tackle payment service fraud. It was further strengthened by the Patriot act in October 2001 and again in 2011 by the introduction of Tax Reporting under President Obama's administration. Furthermore, the various local financial authorities have enacted other rules and policies that the banks have to follow. Some regulations associated with Anti Money Laundering / Sanction Countries that have been in place for many decades:

    • 1970 - Introduction of the Bank Secrecy Act covering the detection and prevention of money laundering
    • 2001 - USA Patriot Act covering the interception and obstruction of terrorism
    • 2011 - Introduction of Tax Reporting by the IRS (USA)
    • Ongoing - New regulations by local Financial Authorities

    Anti Money Laundering (AML)

    There are various challenges in detecting money laundering & Sanctions compliance, not least that the criminal is always looking for new ways to circumvent banks' systems and practices.

    When focusing on AML, we can see criminals will always find new ways to get questionable funds into the financial system. Just like hackers creating new viruses followed by Antivirus providers enhancing their software to combat the latest pattern, the banks are constantly changing their systems to detect and combat the latest developments of criminal gangs. What makes it really hard for the banks is that the procedures to find and fix follow a "rearview mirror" approach. Money launderers create companies, whose only purpose is to launder money. These companies again own a web of companies (often in countries where AML procedures are less effective or the rules are less strict). The ultimate goal is to feed funds into the system without detection and move the money around until it is effectively untraceable. Often family and relatives are part of the system to obfuscate the flow of funds.

    Looking at Sanctions compliance the challenge is a bit different. The banks have to respect such sanctions issued by the local authorities where the account is booked. The biggest two are OFAC the Office of Foreign Assets Control in the US and the European Service for Foreign Policy Instruments. Since sanctions are issued and released again, banks have to do their best efforts in order to respect the different sanctions in force at any time. This makes it very hard for the banks to stay compliant.

    Issues Banks face

    So the big dilemma is setting the right level of risk. Most compliance departments would like to review every significant transaction as it is their job to reduce the risk to the maximum possible amount. This would be prohibitively costly and not practically realistic. At the other end of the scale, the bank could tune the system in a way that only some really awkward transactions are filtered out to be investigated. This, however, leads to a very high risk of noncompliance since the criminal gangs are really smart and find sophisticated ways to bypass the systems.

    The optimal solution lies somewhere in the middle and that's where TIBCO can help.

    So how can TIBCO make a difference in this highly complex world? You can now see why financial crime prevention is challenging by its sheer scale and complexity. If we look at today's detection systems, most of them rely on rules applied to past transactions. Rules have to be defined by experts and follow a clearly identified pattern. There are two main aspects why rules are not sufficient anymore:

    • Usually a specific case has first to be identified before a rule is implemented (remember the comparison with the Virus and the Anti-Virus software)
    • The implementation of rules usually takes a while  techniques like hot deployment are rarely available.

    Another big aspect is that the applications in use today are dedicated special-purpose applications from specialist suppliers. Most users consider these applications to be "black boxes" that are hard to change and expensive to buy and implement. This is where TIBCO technology can help, not by replacing existing applications but by elevating those to the next level by adding Artificial Intelligence and machine learning, paired with real-time Transactional Analytics.

    Benefits in a Nutshell

    • Reduction of "False Positives": By applying artificial intelligence and statistical machine learning models paired with real-time Analytics, the number of False Positives is significantly reduced
    • Business Insights: Business (i.e. Business Risk Management) can access and understand the impact of the model change as the results are visualized
    • Transparency: The Connected Intelligent Platform is fully transparent, and can be enriched with artificial intelligence including having relevant predictors and customized rules being executed in real-time
    • Connectivity: Connects to virtually any data source incl. Big Data clusters and real-time data feeds
    • One-Stop Shop: Covers the full end-to-end process: capture/visualize/analyze/score/processing (integration points covered by adaptors where needed)

    Differentiators of our Connected Intelligent Platform

    • Generic Platform: The platform allows the bank to keep control of each of the different steps in contrast to ?Black Box" Solutions
    • Reusability: Platform to be designed to be applied to many use cases (i.e. Fraud Mgt, customer churn, etc.) resulting in lower cost (license, IT operations, internal know-how)
    • Transparency: Flexibility to develop models using coding or using graphical development catering to different personas of people; tuned to achieve the shortest time to market
    • Scalability: Linearly scalable (horizontally: add servers / vertically: add cores) across all the layers. (Visualization, machine learning, and real-time streaming)
    • Connectivity: Connects to virtually any data source data sources

    Where to go to learn more

    1. A cross-industry AI Streaming Machine Learning tutorial aimed at either business users wanting to set up a machine learning strategy for their business, or advanced data scientists wanting to put their models into production in real time (download or just play). Learn how to adapt it to maverick lot, AML, churn, credit scoring, fraud, and claim management, anytime you need to separate one group of cases from the rest.
    2. A supervised learning template that uses different models at once to allow users to choose the best (download).
    3. A network chart template that allows users to enter relationships between potentially criminal nodes and spot other members of their crime ring. Forensics on potentially huge data volumes (download or just play).

    User Feedback

    Recommended Comments

    There are no comments to display.


×
×
  • Create New...