AI's potential for handling customers' complaints

Synovus branch and image of Microsoft Copilot
Using AI “won't eliminate people’s jobs,” Heather Hajek, director of analytics and compliance management system support at Synovus Financial, said on a panel at the Consumer Bankers Association annual conference. “It helps us do our jobs better."
Microsoft Copilot: Gabby Jones/Bloomberg

Synovus Financial sees promise in how artificial intelligence can aid in its customer complaint management system, but the bank is starting small. 

"I thought it was ironic that a compliance person was going to be up here talking about AI," said Heather Hajek, director of analytics and compliance management system support at Synovus, on a panel about enhancing complaint management through AI and technology at the Consumer Bankers Association's annual conference this week.

Analyzing customer complaints is not just about managing risk, avoiding regulatory scrutiny and identifying systemic issues that would harm the bank's reputation or performance. It can help the bank predict what it can do better in the future to stem consumer attrition and strengthen loyalty. 

"How customers are treated, regardless of the outcome, has significant bearing on their future behavior with the institution," said Brendan Mulvey, a partner at Promontory Financial Group, on a separate panel at the conference about complaint management. 

Tony Antonopoulos, director of compliance at Rockland Trust Company in Rockland, Massachusetts, is in the process of remaking the $19 billion-asset bank's customer complaint program, he said on the panel about enhancing complaint management through AI. The bank will examine both negative and positive feedback. 

As a former examiner with the Federal Deposit Insurance Corp., "I've seen a lot of bad complaint programs," he said.

For him, the good ones have a minimum of five components: clearly defining what the complaint is to employees; easily capturing feedback; aggregating these grievances from both the bank's channels and third-party service providers; analyzing complaints for trends and volume; and reporting trends and data to people who can effect change, including senior and executive management and the board of directors.

The Columbus, Georgia-based Synovus is not yet using AI for logging or reporting complaints from customers, although "in the future, that is something we look to do," said Hajek.

Microsoft, Google, Salesforce, ServiceNow, Blend, Q2 and Intuit are among the software companies that offer copilots. Here's why banks are taking notice.

November 20

Instead, the bank, which has $60 billion of assets, has been using Microsoft's Copilot since last year to kick-start the writing or editing of written letters in response to common consumer complaints. For instance, in a live demo on the panel, Hajek typed "write a customer response letter regarding overdraft fees" into a chat window. The letter that materialized on screen confirmed that the overdraft fee was applied in accordance with the bank's policies, but expressed empathy for the customer, offered to waive the fee as a one-time courtesy and suggested that the customer explore its overdraft protection options with the Synovus team.

The letters that Copilot generates are reviewed by legal and compliance before they are mailed to customers. Synovus is also considering an AI tool within Power BI, Microsoft's data analysis and visualization platform, to summarize executive summaries, trends and elevated risks for its complaints reporting. 

The bank sees such tools as additive. Using AI "won't eliminate people's jobs," Hajek said on the panel. "It helps us do our jobs better. Sometimes you have writer's block. This helps you get started."

The bank is using AI in other areas as well, including in the contact center. A tool from Five9, a cloud contact center software company, listens in on customer calls and summons what the bank calls "knowledge articles," or quick reference guides, for the agents so they can better assist the customer. It also summarizes the calls for agents to review and edit. In the future, the bank hopes to introduce a scoring system that assesses how well an agent handles certain topics and what they need to improve in. This could help the bank determine whether more knowledge articles or training are necessary. 

Synovus has also found one lower-tech way to improve the quality of complaint reporting: giving examples to front-line staff about how their complaint reporting has resulted in the bank identifying systemic concerns and improving products or services.

"We have been able to demonstrate how their reporting of complaints in turn helped them do their jobs better," she said.

For reprint and licensing requests for this article, click here.
Artificial intelligence Bank technology Technology
MORE FROM AMERICAN BANKER