A prominent financial institution in the United States has significantly alleviated its sanctions compliance workload following the implementation of AI-driven agents from SymphonyAI, resulting in a 90% reduction in manual tasks and an impressive tenfold decrease in alert review times.
The institution faced challenges due to an overloaded screening infrastructure. Its transaction monitoring system relied on broad name-matching methodologies that frequently resulted in high false-positive rates, which in turn overwhelmed compliance teams with increasing alert backlogs, according to SymphonyAI.
Previously, each sanctions alert required extensive manual investigation. This involved reviewing transaction details, conducting entity background checks, examining historical records, and performing web research, a process that could take over 100 minutes for a single case.
Furthermore, analysts were tasked with reviewing numerous media articles to determine if flagged content pertained to the appropriate individual or entity, adding to the strain on their already taxed resources.
To tackle these issues, SymphonyAI initiated a proof of concept (PoC) to showcase how its Symphony Risk Intelligence (SRI) platform, and specifically its SRI Agents, could streamline much of the labor-intensive process.
The agents were customized to operate based on the client’s own policies and procedures, enabling them to perform a variety of tasks autonomously. These tasks included automated entity resolution, where agents assessed transaction participants—such as senders and receivers—by gathering publicly available information and conducting thorough analyses.
The agents also engaged in name disambiguation, relationship assessment, and background verification to identify false positives, while flagging cases that necessitated human review, complete with detailed justifications adhering to the client’s internal guidelines.
The outcomes from the PoC were remarkable. The deployment led to a 99% reduction in false positives by independently managing and filtering sanctions hits, along with a 90% cut in manual efforts. Average alert review times plummeted by a factor of ten, and agreement between agent and investigator decisions exceeded 98% of the time.
For scalability, 50 SRI Agents operated concurrently, processing over 300 alerts each hour, thereby conserving thousands of hours for investigators. Tasks that previously took days were now accomplished within minutes, with each decision bolstered by comprehensive subject research, linked resources, and clearly stated policy rationale to aid investigators.
The financial institution intends to expand the application of SRI Agents across additional compliance workflows, aiming for what SymphonyAI terms “Always-on Compliance.”