Skip to content

The Closer you Look, the Less you See!

Unlocked

For those of you who haven't watched the movie "Now You See Me", it's a story about a group of magicians robbing a bank - while performing a show in Las Vegas! It was natural for me to juxtapose it with the world of financial crime.

For those of you who haven't watched the movie "Now You See Me", it's a story about a group of magicians robbing a bank - while performing a show in Las Vegas! It was natural for me to juxtapose it with the world of financial crime.

Think about the quotes in the movie - "Nothing's ever locked' and 'the closer you look the lesser you see’.'The eye' of organized crime is where some very talented individuals, albeit with warped ambitions, strive to be. Many of them are insiders, so it will be a continuous chase for fraud prevention professionals & cops alike as they keep inventing new tactics.
 
To stay on top of their tricks we need to have a few 'up our sleeves'. Text analytics is a simple and effective technique that can be used to mine meaningful information - categories, sentiments and so on. The National Association of Certified Fraud Examiners (NACFE) estimates that more than 80% of fraud within an organization can be identified from analyzing text based data. In addition to internal data like emails, documents and spreadsheets, the same technique needs to be applied to something we are all part of - Social Media.
 
A few weeks earlier, I attended a financial crime management and security conference in Singapore, where I heard a rather interesting story. A customer at the ATM of a well known multinational bank discovered she was able to withdraw more cash than her account balance included and being a kind woman that she was, posted about it on Facebook. Soon, queues of opportunistic customers had formed outside many of these ATMs. While the bank recovered some of the money later, it took a long time for the bank to discover this was happening. Text mining and analytics off the social media streams in this case could have been an opportunity for the bank to prevent this from happening.
 
Such information becomes relevant to a Bank's fraud prevention teams to uncover rings of organized fraud using the simple Markov assumption of Homophily or guilt by association. In the ideal world this is not limited to search or visual link analysis as a forensic tool. It extends to a event triggered scoring mechanism akin to the payment transaction scoring we are all used to. With Complex Event Processing (CEP), it is possible to predict the occurrence of an event (fraud) by creating event correlations using text analytics on multiple data from varied sources, including social media feeds. It is indeed complex, but seems that is the next generation of fraud detection technology!