Michigan State University Federal Credit Union moved its client-facing chatbot, Fran, in front of its live chat agents in May, and the chatbot is now fielding 73% of the bank’s chat inquiries.

“When we moved Fran in front of our live chat agents, two things happened,” Ami Iceman-Haueter, chief research and digital experience officer at the $7.45 billion MSUFCU, told Bank Automation News: The chatbot could answer questions that typically wouldn’t have gone through it and inquiries directed from the chatbot to an agent led to more meaningful conversations with members dealing with more complex issues.

Fran has a 98% accuracy rate for the questions it fields, Iceman-Haueter said.

The bot can answer questions regarding account-specific information, including payment details; unlock credit cards; find account numbers; and provide loan rates and can also walk members through step-by-step solutions for problem-solving, according to the MSUFCU website.

Additionally, the chatbot is doing the work of 34 full-time employees, up from 21 full-time employees in March, Iceman-Haueter said. The credit union has not downsized its team at all; the technology is taking over roles that were difficult to fill amid workforce challenges during the past several years.

“We have no intention of downsizing in any capacity, if anything we’re really trying to narrow in on what next service things we can provide to members,” she said.

Building Fran

MSUFCU teamed up with Boost.ai in 2021 to reintroduce the client-facing chatbot and launch its own internal chatbot, Gene, Iceman-Haueter said.

“Both bots now operate with [Boost.ai] and they’re continuing to help us partner and grow those platforms,” she said.

For example, the chatbots go through quarterly updates to ensure that they address client and internal needs, she added. “We’re also continuing to roadmap out what [the chatbot] expectations will be in 2024 and 2025 to make sure that we can continue to meet our member support.”