A major US financial institution has dramatically reduced the burden of sanctions compliance after deploying AI-powered agents from SymphonyAI, achieving a 90% reduction in manual effort and a tenfold improvement in alert review times.
The organisation had been struggling under the weight of its own screening infrastructure. Its transaction monitoring system relied on broad name-matching logic that generated persistently high false-positive rates, leaving compliance teams overwhelmed by a growing backlog of alerts, SymphonyAI said.
Each sanctions hit required investigators to manually sift through transaction details, entity background checks, historical records, web research, and Requests for Information (RFIs), a process that could consume more than 100 minutes per case.
On top of this, analysts were expected to review potentially dozens of media articles for each hit to confirm whether flagged content genuinely related to the correct individual or entity, adding further strain to already stretched teams.
To address these challenges, SymphonyAI conducted a proof of concept (PoC) demonstrating how its Symphony Risk Intelligence (SRI) platform and specifically its SRI Agents capability could automate much of this labour-intensive workflow.
The agents were trained on the client’s own policies and procedures, enabling them to carry out a range of tasks independently. These included automated entity resolution, in which agents analysed transaction participants such as senders, receivers, and sanctions matches by gathering and assessing publicly available information. The agents also performed name disambiguation, relationship analysis, and background verification to identify false positives, while flagging cases that required human escalation, complete with detailed justifications aligned to the client’s internal guidelines.
The results of the PoC were striking. The deployment achieved a 99% reduction in false positives by autonomously matching and discounting sanctions hits, alongside a 90% reduction in manual effort. Average alert review times fell by a factor of ten. In terms of accuracy, agent and investigator adjudication decisions were in agreement in more than 98% of cases.
To demonstrate scalability, 50 SRI Agents were run in parallel, adjudicating more than 300 alerts per hour and saving thousands of investigator hours. Tasks that had previously taken days were completed in minutes, with each decision supported by detailed subject research, linked sources, and clearly outlined policy rationale to assist investigators in their assessments, it said.
The institution is now looking to extend the use of SRI Agents across a broader range of compliance workflows, with an eye towards achieving what SymphonyAI describes as Always-on Compliance.
For more insights, read the full story here.
Copyright © 2026 FinTech Global
