Flywire Adds Machine Learning Capabilities to Its Payment Platform

Optimizes FX, Speeds Receivable Matching, and Reduces Fraud Involved with Diverse Cross-Border Payments for Growing Global Customer Base

Flywire, a company that solves complex payment problems for leading businesses and institutions, announced the addition of powerful new machine learning capabilities to its cross-border payment and receivables platform. The enhancements improve the payment-to-settlement time, increase security, and reduce costs for both payers and receivers by further automating and streamlining reconciliation of the growing number of international payments coming from different countries in different currencies.

There is significant interest in expanding the application of artificial intelligence and machine learning in the financial services sector. Research advisory firm Autonomous NEXT estimates that financial firms globally can eliminate up to 20% of costs through the implementation of the technology while also improving service quality. In the area of middle office processes such as payments, the company’s analysts point to increasingly complex regulations and real-time processes which are making artificially intelligent oversight, risk-management and KYC systems very valuable.

Read More: PayPal’s First Blockchain Investment Is an Identity Ownership-Driven Start-Up

Typical legacy payment platforms employ rigid, rules-based systems to perform ‘best effort’ reconciliation of invoices with monies received for businesses and institutions collecting payments. These platforms are limited in their ability to support the ever-evolving business requirements, multiple currencies, and myriad payment methods involved in collecting cross-border transactions. As a result, a significant manual effort is required to review transaction records and reconcile payments.

With the addition of machine learning-enabled deep neural networks and reinforcement learning techniques, Flywire has enhanced its ability to streamline the identification and reconciliation of complex, cross-border payments in real-time. Users will be able to automate the matching of 90% or more of their cross-border transactions and gain additional improvements as the models learn. Furthermore, the machine learning algorithms require minimal supervision to learn and support new payment methods and can confirm payment sources, detect anomalous payments, and escalate these to Flywire’s Compliance and Operations teams for review. The new capabilities also further optimize FX conversion.

Read More: 4 Steps to Optimize Cloud Spend: How to Address 2019’s Top Cloud Initiative

“Accepting payments across borders is a highly complex process that increases the cost of collecting monies, opens up FX and fraud risks, and requires enormous operational investment. The result is poor cash flow, higher costs and unsatisfied customers,” said Jason Moens, VP of Product at Flywire. “For the last ten years, we’ve tackled the inefficiencies in global payments by connecting all the entities involved to make them faster, more secure, less expensive, and more transparent for all.

“The addition of advanced machine learning capabilities further streamlines our clients’ payment and receivable operations and removes more of the potential risks that can negatively impact fundamental parts of their business. This allows them to offer customized payment solutions to more of their customers — wherever they are in the world.”

Read More: Hyperledger Welcomes Nine New Members to Its Expanding Enterprise Blockchain Community