Innovation Description: Digitized payments are playing a role in economic transformation of the world population with inclusive services. This innovation is being driven by the remittances market whose traditionally high prices are a major pain point for economic inclusion.

 

As fintechs and banks monitor trillions of transactions to fulfill regulatory requirements, doing the task manually using human labor is becoming impossible, especially with continuous increase at a rate of between 13% to 25% globally with emerging markets leading the growth especially in digital wallets.

 

The old approach would be to build a model and catch the bad guys based on past patterns of behavior. But this is not working anymore, as financial crime vectors can change in a matter of seconds.

 

ThetaRay’s AI and machine learning technology, a methodology called “artificial intelligence intuition,” uses proprietary unsupervised and semi-supervised machine learning algorithms that replace human bias. This empowers the system with an unmatched ability to recognize anomalies and find unknowns outside of normal behavior, compared with rules-based or supervised machine learning solutions.

 

The ThetaRay machine-learning (ML) methodology and algorithms are applied to dozens of risk indicators associated with financial crimes. This risk-based AI approach paints a clear picture for compliance teams and enables them to detect abnormal activity within large sets of data and to effectively calculate and pinpoint transactions indicating suspicious activities and prevent blocking and slowing down transactions of ordinary people.

 

ThetaRay’s ML process also provides insights on customer identity in cases with limited visibility where KYC information is lacking, creating risk profiles of non-customers and transparency across complex paths.

 

Using 19 patented algorithms, the system identifies abnormal customer or transactional behavior patterns which deviate from the expected or “normal” behaviors from an AML risk perspective. This enables the rapid discovery of both known and unknown money laundering threats, with a 99% reduction in false positives compared to rules-based solutions.