Refine Intelligence

Fraud Solution Profile

Banks are flooded with more AML alerts than ever, but 95% of them are false positives triggered by legitimate activity like purchasing real estate, paying college tuition, doing renovation work or receiving wedding gifts. Once upon a time, bankers knew their customers well enough to resolve many of these alerts based on their knowledge of each account holder’s life events. But that era has unfortunately passed, and instead, bankers are caught in the middle between AML teams and their annoyed account-holders who have to answer invasive questions about their money. 

Refine Intelligence comes to reset AML practices for the 21st century, handing back the superpower of knowing your customer well enough to understand the reason behind changes in their financial activity. It’s based on identifying the “good guys” by quickly resolving false positives, rather than attempting to catch the small minority of “bad guys.” 

Refine Intelligence offers two products to stop the frustration and speed up resolution, thereby removing friction from the customer relationship, cutting churn rates, and freeing bankers to concentrate on revenue-driving activities. 

Digital User Outreach 

An automated data collection process that lets AML investigators reach out to customers via mobile push, email or SMS. Instead of weeks of back-and-forth emails with the branch, it takes customers about 60 seconds to complete these digital “requests for information,” and they can do so at their own convenience without being interrupted by a phone call, which boosts completion rates to 85%. The process offloads over 70% of branch or call-center AML-related work.

The AML investigation team receives a real-time explanation of the nature of the alerted activity and the source of funds or relationship with the beneficiary in a clear audit trail. Refine’s process uses consistent, structured questions that avoid bias and tipping-off risks. 

Life Story Analytics

An AI-driven database that creates a baseline of good behavior, based on millions of financial records, and then maps each transaction alert to its most likely life stories that are ranked by probability. The solution taps into third-party API checks, open-source data, and the user’s transaction history to ascertain whether the flagged transaction is a false positive or genuine fraud risk, and maps it against the baseline to identify life stories that could explain it.