When a financial services company suspects a financial crime is taking place, they legally must submit the information to the US Treasury as part of the Banking Secrecy Act (BSA) through FinCEN (Finance Crimes Enforcement Network). In today’s technology-driven world, this is a Herculean task, with millions of records requiring attention, and as banking, financial services, and insurance(BFSI) have shifted to digital commerce, digital financial crimes have surged.
To help manage this process, one of the largest financial institutions in the world turned to Arria’s Natural Language Generation (NLG) technology to expedite the reporting of Suspicious Activity Reports (SAR). Instead of manually analyzing data and writing a report, now, when a suspicious event occurs, an investigator compiles data on the event, sends it through Arria, and receives a comprehensive, auto-generated report.
The impact of NLG on a financial institution
Arria NLG gives investigators superhuman powers by reducing report creation time by 80%. The institution was creating 400,000 reports annually. With the shift to Arria, they saved approximately 90,000 employee hours per year, and the automation guaranteed accuracy.
Using the Arria NLG platform, the institution:
- Maintains compliance with BSA financial crimes reporting requirements
- Eliminates backlogs by automating reporting in real-time
- Shuts down malicious actors quickly
- Frees investigator time with reporting efficiencies
An example of an NLG narrative for SAR reporting
Arria’s natural-language narratives are customized to the industry terminology and can also be customized to the voice of a firm so that the language produced is natural and reads as if written by an analyst in the organization. Here is an example:
MAG Bank, Inc. (MAGBk) is filing a Suspicious Activity Report (SAR) due to an activity indicative of payment card fraud conducted through MAGBk’s point-of-sale (known as “Seller”).
This investigation reviewed all activity from 1st January 2012 through to 31st March 2022. The review identified 26 suspicious transactions totaling US$117,314.10, conducted across five MAGBk accounts from 2nd June 2012 to 25th June 2019.
Of the 26 transactions, 6 were declined, and 20 were successfully processed, totaling US$129,673.40. The majority of the transactions were conducted by card-not-present. Card-not-present is an activity made through electronic invoices in which a customer inputs their payment card information directly into an invoice they received via email or via text SMS.
Note that MAGBk later received four chargebacks totaling $12,359.26, for fraud-related reason codes.
According to MAGBk records, Mrs. Fictionale opened account number A1234 on 2nd June 2012, Mr. Smith opened account number L9876 on 5th November 2014, Doe opened account number R5432 on 9th April 2017, Mrs. Jones opened account number Y3456 on 27th November 2018, and Mr. Bojangles opened account number C45678 on 25th June 2019 to process payments for various types of businesses.
A review was prompted by an internal alert. Of the 56 transactions, six were declined, and 50 were successfully processed, totaling $138,183.79. The majority of the transactions were conducted card-not-present, through electronic invoices in which a customer inputs their payment card information directly into an invoice they received via email. MAGBk later received four chargebacks totaling $12,359.26, for fraud-related reason codes.
Arria’s long experience with BSFI
Arria is the global leader in natural language generation technology and has been working with the banking, financial services, and insurance (BSFI) sector for years. Our experience, reputation, and ease of integration have meant that many of the world’s leading BSFI organizations trust Arria to help them with SAR and other reporting.
To learn more about how Arria NLG can automate your organization’s reporting, download the Forrester TEI report.
Learn more about how Arria NLG helps a wide variety of industries.