Verafin uses fuzzy logic and AI to fight bank fraud and money laundering

Bank fraud and money laundering are big business for criminals and big headaches for financial institutions. Fraud losses run into billions of dollars every year. Verafin is putting a dent in those loss figures with its consolidated anti-fraud and anti-money laundering solution. Fuzzy logic, artificial intelligence and broad databases are the keys to detecting and often preventing malfeasance.

In November I wrote an article about how payment card skimming is on the rise (see "On Santa's 'Naughty' list: credit card skimmers"). Skimming is the illegal act of copying the information stored on the magnetic stripe of a credit or debit card using a small device called a skimmer. Thieves then imprint the stolen data onto a blank card to create a replica that can be used virtually anywhere the legitimate card would be accepted, such as at an ATM or point-of-sale terminal.

Financial institutions (FIs) that issue cards are the ones that suffer the majority of the losses. According to the American Bankers Association, U.S. industry losses from debit card fraud were estimated at $788 million in 2008. Unfortunately, skimming is just one form of fraud and other financial crimes that FIs face. They must also grapple with stolen, forged or altered checks; wire fraud; money laundering; phishing and Internet fraud; internal theft; and many other types of transactional crimes. There are no hard statistics on the collective total value of losses, but experts say it's well into the range of tens of billions of dollars each year.

IN PICTURES: Credit card skimming: How thieves can steal your card info without you knowing it

It should come as no surprise, then, that FIs such as banks and credit unions are fighting back. In fact, both U.S. law and international cooperative agencies require that FIs implement strong anti-fraud and anti-money laundering (AML) programs, which can be either manual or automated with technology. It's not just about protecting consumers and preventing financial losses for the FIs. With a global concern about terrorism and organized crime such as drug trafficking, anti-fraud and AML programs are key to shutting down the movement of money for illegal purposes.

In the complex world of financial transactions, automated solutions are essential for connecting the dots that humans might miss. One technology company in particular is a real standout in the banking world when it comes to detecting and preventing fraud and money laundering. The company has an interesting background that helps with the creative application of technology toward the financial industry.

Verafin Inc. was founded in 2003 by a trio of post-graduate engineering students. They met in a university robotics lab where they were designing navigation systems for robotic vehicles. An investor approached them and asked if the engineers could apply the advanced technologies from robotics to financial services to address the problems of money laundering, fraud and other financial crimes. According to Jamie King, co-founder, president and CEO of Verafin, the two fields are similar in that they both have a lot of transactional data. By applying algorithms, it's possible to make sense of the data -- whether you are directing a vehicle to swerve to avoid an obstacle or you are detecting the possible malfeasant use of a debit card.

Most of the anti-fraud solutions that existed at the time were fairly simplistic, rule-based systems. Structuring the rules can be cumbersome and difficult to manage, and it's virtually impossible to create a rule for every situation. What's more, rules can only look at customer activity and banking systems in a very siloed way. For example, there are many ways a customer can make a bank transaction. He can make a cash deposit, cash a check, make a credit purchase, and so on. Most rule-based systems are limited to looking at one type of a customer's transactions at a time, but not all of them holistically. Without a holistic view of what's happening, a lot can be missed.

Instead of rules, Verafin's solution applies fuzzy logic and artificial intelligence in its algorithms. The vendor also uses a vastly broader set of data points in order to take a holistic view of banking transactions. Verafin is one of the first companies to merge fraud and anti-money laundering into one solution, coining the word "FRAML" to indicate the integrated nature of what are usually two separate solutions. Verafin looks at customer behaviors as a whole across the bank. This means the solution can look at the size and types of a customer's transactions, the times when they occur, the velocity of the transactions, and where they occur.

Importantly, Verafin also incorporates a profile of the customer into the data analysis to see if the transactions are out of character for that specific customer. For example, if an elderly person's debit card is being used late at night in such a way that he's never used it before, Verafin will flag the transaction as a potential fraud case. According to King, this scenario could be indicative of elder abuse. This same kind of analysis could potentially prevent the fraudulent use of a counterfeit debit card at the ATM.

Another shortcoming of rules-based systems is that they use hard thresholds. By applying fuzzy logic, Verafin gets away from hard thresholds and has the flexibility to consider the "gray areas" that may or may not indicate fraud. For example, if you create a rule dealing with a large transaction, at some point you have to define what "large" is. So, if $500 is "large," then with a traditional logic system, $499.99 is "not large." With fuzzy logic, you can say that $200 is "small" and $600 is "large," and everything in between is gradually increasing from small to large. This gives a bank more flexibility to judge the risk of a transaction and the likelihood that it is fraudulent.

In addition to using a bank's transactional and customer data, Verafin regularly ingests data from the federal government, including "314A" lists of known terrorists or criminals. Using such a wide range of data as input to Verafin's analysis yields much more effective results.

Verafin's solution is a feature-rich complement to core banking systems. The solution can run in the cloud, hosted by Verafin, or run in a bank's or credit union's own data center. Much of the data analysis is performed offline overnight, but some features operate in real time, such as validity checks on ATM transactions. Verafin generates alerts when there is suspicious activity, and sends reports as needed or required to the Financial Crimes Enforcement Network (FinCEN). The solution also has a fully functional case management system to keep track of cases of suspicious activity.

While Verafin can't disclose much about its customer base, King says the solution is used today by more than 700 banks and credit unions in North America, the majority of which are in the United States. With a 99% customer retention rate, this company must be producing the results the FIs are looking for.

Linda Musthaler is a principal analyst with Essential Solutions Corporation. You can write to her at

About Essential Solutions Corp: Essential Solutions researches the practical value of information technology, and how it can make individual workers and entire organizations more productive. Essential Solutions offers consulting services to computer industry and corporate clients to help define and fulfill the potential of IT.

Learn more about this topic

Credit card skimming: How thieves can steal your card info without you knowing it

ATM skimming: How to recognize card fraud

Fraudsters find holes in debit card fraud detection

Join the Network World communities on Facebook and LinkedIn to comment on topics that are top of mind.
Must read: 10 new UI features coming to Windows 10