Finastra is seeing high adoption for its OperatorAssist, an AI-powered tool designed to help bank operators resolve payment exceptions faster. 

The tool, launched in March, addresses one of the costliest problems in payments processing: investigation of failed transactions and manual resolution, Mihail Duta, director of global solutions consultant and payments at Finastra, told FinAi News. 

Nearly 98% of the payments go through STP (straight through processing) but the remaining 2% of the payments might get reviewed by the bank’s payments operators due to some discrepancies, Radha Suvarna, chief product officer for payments at Finastra, told FinAi News. 

“The transaction sometimes could be [labeled a] mismatch in the destination country or in a domestic transaction, some [mismatches occur] in the account number, routing number, name, sanction screening,” Suvarna said. 

Inefficient exception and investigation handling costs the payments industry around $2 billion each year and can affect up to 700 million transactions annually, Duta said.  

The AI-agnostic tool can reduce manual investigation time by 20% to 30% on average, saving users more than 1.5 hours daily for a payments operator, Duta said.

Finastra research shows “23% of our commercial clients rank payment errors as their biggest pain point with their current payments provider,” he said. 

The tech 

OperatorAssist functions as a co-pilot rather than an autonomous decision-maker, using multiple large language models depending on the use case, Duta said, adding that the tool was in beta testing for a few months. 

The tool can surface payments through natural-language queries — for example, pulling all payments in repair from a week ago under $1 million — and recommend the top five resolution options when an operator encounters an error message, such as a mistyped account number, he said. 

“We’re not embedding AI in payment processing. We’re just enabling the operator to make quicker, faster, more efficient decisions,” Duta said. 

The efficiency gains can be substantial. The tool can compress exception handling “from minutes to seconds,” depending on the complexity of the scenario, he said. 

OperatorAssist is the first iteration of a longer AI roadmap at Finastra and the launch coincides with what Duta describes as a shift in the industry’s posture toward AI.  

AI in payments “is moving from an experimentation stage to a mandatory infrastructure,” he said, adding that fraud and AML are among the top categories where AI meets payments. 

FIs that work with Finastra include:

• ING Bank;

• Beacon Credit Union;

• Cache Valley Bank; and

• Lloyds Bank. 

Register here for the FinAi Lending Summit, set for Oct. 7-8 in Las Vegas.