From Building Minebots to Digging for Dirty Money

By John Tozzi

Jamie King's attempt to build mining robots turned into software that helps banks find money launderers in a sea of financial data

In 2003, Jamie King was trying to develop machines that could replace humans in dangerous nickel and iron-ore mines. Then a robotics PhD student at Memorial University in Newfoundland, Canada, he used artificial intelligence to interpret the vast data streaming from the robots' sensors, which measured the distance of objects in every direction a thousand times a second. Although mining companies passed on the idea, King later discovered an unlikely new use for his minebot software: catching money launderers.

Banks collect huge data sets on the millions of transactions that move money in and out of their accounts. When university colleagues introduced King to a financial-software entrepreneur, he got a peek at such a pile of numbers and realized it looked "very much like a whole bunch of sensor readings." King co-founded Verafin in 2003 to sift bank data for patterns that could indicate fraud, drug trafficking, or terrorist financing.

At the time most banks used rules-based software that flags transactions if they match a pattern defined beforehand as suspicious, such as several transfers of money overseas in a short period. Verafin compares activity to an account holder's profile and past behavior and assigns each transaction a "probability score" that represents the likelihood it's legitimate. King says that evaluating probabilities lets Verafin discern suspicious patterns that slip through conventional systems. Say a drug lord in Chicago needs to get cash to Mexico. He might make wire transfers and deposits to multiple accounts, buy prepaid debit cards, and send payments to a shell company off-shore. By itself, any one of these actions might look perfectly fine. Verafin identifies the extremely low probability of such a pattern occurring innocuously, says King.

Verafin had C$9 million ($9.3 million) in revenue in 2009, the latest figures King will share. More than 700 banks and credit unions use Verafin's software, most with between $100 million and $20 billion in assets. Even relatively smaller banks, like Beach Business Bank in Manhattan Beach, Calif., with assets of $300 million, have to track "money that's moving around quickly and electronically, where you don't have as much information about where it's ending up," says Melissa Rickabaugh, the bank's anti-money-laundering compliance officer. She says Verafin will automatically flag an account when large deposits are followed quickly by wire transfers out, which the bank's previous software didn't do.

A native of Newfoundland, King describes himself as "an inventor who likes tackling hard problems," which is what first drew him to robotics navigation. Sniffing out money laundering has proven equally tough, he says: "It's always a challenge to stay ahead of the criminals."

EDUCATION

Left a robotics PhD program in 2003 to start Verafin.

TECHNOLOGY

Software combs financial transactions for suspicious patterns.

MARKET

700 institutions use it to comply with money-laundering laws.

Source: Business Week