A global bank had a problem: they had to identify people and organizations who could not be part of any transaction that involved moving money in or out of the bank.

When the government places sanctions on people and organizations, it’s the bank who must ensure that those sanctions are respected and that the money doesn’t move. There are many complicating factors. Names are not always spelled correctly or spelled the same way on different lists and documents. Some entities might even deliberately misspell their names to circumvent sanctions. Transactions can occur on a variety of systems, including ones that process direct transfers, wire transfers, and SWIFT. Some transactions are batch, some are real-time.

When the transaction is initiated, the bank must determine if any of the parties to the transaction are proscribed. If so, they must immediately halt the transaction or at least refer it to someone who can review it. The bank needs to act quickly, so that real-time transactions aren’t interrupted. The bank also needs to avoid false positives (blocking legitimate transactions involving parties with names similar to the names of proscribed parties).

The standard software solutions might have been fine for a smaller bank. For a global bank faced with thousands of transactions per second, such solutions just couldn’t keep up. The bank tried writing their own software, but its matching capabilities were limited. The regulators were not happy.

Fortunately, the bank came to Ab Initio for help.

For Ab Initio, processing thousands of transactions per second with incredibly low latency is no big deal. The bank took full advantage of Ab Initio’s incredible speed, scalability, and powerful matching capabilities, to process transactions as they came in. Using Ab Initio’s comprehensive, rules-based pattern-matching technology, the bank could rapidly develop rules that analyzed millions of records and quickly determine the likelihood that “this name is that name.” They could define the threshold at which two names can be assumed to be the same name, and easily test to verify that the results made sense.

The bank ended up feeding their transaction stream through their original software and through Ab Initio’s pattern-matching software. The two results were consolidated into one answer that the bank could be confident in. With this new approach, the bank can now identify sanctioned individuals and groups with remarkable accuracy while minimizing false positives and maintaining extremely low latency. The bank is happy. Even more importantly, the regulators are happy.