Choose language:
Pratite nas:
back

INFIGO IS at F2 - Future of Fintech

15.11.2017

September 14, 2017., Kaptol Boutique Cinema & Bar in Zagreb was teeming with bankers and other financial experts. All of them flocked to the sold-out conference F2 - Future of Fintech to discuss whole range of technical questions and burning trends in financial sector

In track "Security and regulations" (there were also "Digitalization of financial markets" and "Financial innovations") was our own Sanja Zeljković, fraud consultant at Infigo IS, with her lecture titled "Cutting edge FRAML". She talked about the acts of fraud and money laundering, with a special focus on similarities and dissimilarities between those two activities. Of course, she showed how to fight both of them using FRAML system.

As Sanja said: "Fraud and money laundering are often two sides of the same coin so using a holistic approach is the logical step. And that holistic approach is FRAML".

Why FRAML?

Firstly, financial fraud is an attempt to get money in unlawful way - that is possible through different financial channels (ATMs, Internet banking, Point of Sale...), products (exploiting a whole range of accounts and automated services), and transactions (deposits, withdrawals, transfers). And fraud usually happens through counterfeiting, theft, and impersonation.

On the other hand, money laundering is an attempt to mask the source, and the sum, of money, and it is used from tax avoidance to terrorism funding.

The cold hard fact is that solving the first, and the second, problem doesn't get any bonus points to financial institutions  - bottom line is, financial institutions wouldn't be at a loss if they allowed fraud and/or money laundering to exist. But if permitted, it would ruin the trust in financial system and led towards destabilization of modern society as we know it. And nobody wants that to happen.

FRAML, FRaud detection Anti-Money Laundering, is an attempt to put the fight against fraud and money laundering under one system. On financial side, two systems under one roof, is more efficient use of resources. On practical side, a lot of fraud ends up with money laundering so it is logical to fight against both of them in one place, and it is also important to see how this two criminal acts impacts the whole financial institution.

FRAML systems often lead to better customer understanding since fraud and money laundering use the same channels as other financial transactions by legitimate users. And as FRAML monitors all, it could be used as a source for great metrics.

How FRAML works?

FRAML systems could be split in two camps - first one is rule based and the other is behavior based. Rule based system includes using hundreds, if not thousands, of rules that search for suspicious activities. But a good rule based system has to have a great understanding of human nature; for example it is not necessary a fraud if customer spends one month a thousand euros while for the past 14 months he spent only 400. He might have just bought a new TV! In essence rule based system is just a complex jungle of IF, AND, THEN triggers.

Behavior based systems is based on fuzzy logic, artificial intelligence, and machine learning. While in a rule based world the answer is "yes" or "no", fuzzy logic enables us to have an answer that is somewhere between. So the system is basically telling us, "I think there is a 68 percent chance this is fraud". In simple terms, system is more flexible. Artificial intelligence, and machine learning enables the system to adapt while searching to new patterns.

And now, which one is better? We at Infigo think that both of them have certain merits so we took the best of both and made our own hybrid system that has been in development for the last five years - and we aren't the only ones who think our system is great; our clients, regional banks who are using our FRAML, are quite happy with the results. FRAML is a combination with many gears that work in perfect harmony, and luck will have it, Infigo IS just happens to have one of those based on Splunk. You can read more on it in this brochure