Giant Oak, a specialist in the application of artificial intelligence and machine learning for the regulatory and security needs of business and government, has announced new updates to its GOST® product. The Spring 2021 GOST Release enhances performance of entity resolution, improving efficiency and effectiveness in the screening process. Additionally, it introduces new search processing infrastructure to support large-scale, ongoing monitoring on a near-real-time basis. Supported by a strategic investment from one of the largest global investment management firms, the new GOST updates can quadruple search speed for a customer, meaning enterprise users can run just shy of one million searches per day, per use case across GOST’s billions of data sources. The combined improvements to scale and entity resolution are designed to allow GOST users to keep a pulse on large populations while handling an efficient, effective, and high-quality set of leads at a lower overall cost.
“We’re constantly working to improve and adapt our technology to empower our champions, the communities of GOST users,” said Gary M. Shiffman, PhD, founder and CEO of Giant Oak. “With these GOST updates, people now have a powerful machine-learning-based tool to easily search and identify the most important information on the right people.”
In December 2020, Congress approved the AML Reform Act – new legislation requiring financial institutions to implement efficient and effective risk-based approaches to countering money laundering and other financial crimes. To comply with the new rules, banks and regulators must prioritize continuous screening and vetting, and look across all available data, to include publicly available information.
Many screening and vetting products on the market today match names on individual artifacts to produce a result, with no aggregate measure of entity-level resolution across artifacts. Giant Oak says this method yields upwards of 99% false positive rates in large enterprise monitoring populations. Banks, for example, invest heavily in third-party professional services to manually reduce false positives. Without granular control over thresholds (i.e., precision and recall), the volume of busywork becomes expensive and unmanageable.
“We look forward to these new updates exceeding client expectations and ensuring a faster and more efficient screening process,” said Harsh Pandya, president of Giant Oak. “The backing from a multinational financial services company was crucial in enabling these developments. In most tools on the market, matches are limited to structured data – usually, lists – meaning the user will miss a lot of risk-relevant information. With the updates available to customers in the Spring 2021 Release, we can reduce the false positive rate and find more true positives at a lower overall cost in nearly all adverse media screening use cases.”