Case study

US credit union automates up to 2,000 internal interactions per month with conversational AI

MSUFCU

MSU Federal Credit Union developed and launched a virtual agent pilot project in only 10 days to provide critical internal support for customer-facing employees

The relationship between a credit union and its members is crucial to its success as a financial institution. Unlike traditional banks, credit unions are cooperative organizations that focus on delivering consistently exceptional service to their members. So, it's no surprise then that they have earned a reputation for being the go-to choice in banking for consumers who want a more personalized touch. Credit union members expect a great customer experience when it comes to dealing with their finances - whether it's in-branch, over the phone or via direct messaging channels.

MSU Federal Credit Union is the largest university-based credit union in the world with over 300,000 members. Founded in Michigan in 1937 to provide members with financial success and stability following the Great Depression, it now has 21 branches across the United States and employs more than 900 people.

MSUFCU's results at a glance...

FAST: Virtual agent launched in 10 days
EASY: 100% employee approval after 4 weeks
SCALABLE: 2,000 employee-to-employee interactions automated per month

In 2020, MSUFCU selected boost.ai to pilot a virtual agent project aimed at assisting the credit union's employees to serve its members more efficiently. The goal was to provide customer support staff with a user-friendly interface for finding answers to support-related questions. Instead of needing to put members on hold to check with a manager or call the internal support line, employees would be able to query the virtual agent in real-time and have access to consistent, accurate information at their fingertips.

"We've moved to a universal model where our employees now serve more like advisors," said Benjamin Maxim, Assistant Vice President of Digital Strategy and Innovation for the credit union. "We initially looked at conversational AI as a way to extend our service hours, but quickly realized its potential to streamline our internal support processes and make our employee's lives even easier."

10 days to launch and impressive results in just 4 weeks

Working with boost.ai's conversational AI platform, Maxim and his team trained their virtual agent to answer questions on the credit union's most popular knowledge base topics in a 10-day build-out. The pilot project was shared amongst a group of 60 employees across the company's branches, call center and live chat teams. They were tasked with using the AI chatbot to help them assist customers over a period of four weeks and were surveyed throughout the process.

It was essential for Maxim to get the artificially intelligent chatbot technology in front of users as quickly as possible. The more engagement the virtual agent had with human agents, the more feedback and insight his team gained into how to improve the overall customer experience. "What's absolutely critical with chatbot projects, is that you need people to interact with them to know if they're going to work well," added Maxim. "If you spend all this time working on a project and you don't get it in front of your support team, you won't ever know if it succeeds."

“Boost.ai’s user-friendly software makes it easy to analyze conversation data and use it to make the chatbot more functional.”

— Benjamin Maxim, Assistant Vice President of Digital Strategy, MSU Federal Credit Union

The results were overwhelmingly positive, with staff scoring the virtual agent increasingly higher as the weeks progressed. From week 1 to week 4, answers to the question 'How beneficial would the chatbot be for day-to-day work?' improved from 50% stating it was 'Very' or 'Moderately' beneficial, to 100% by the end of the pilot period. The question 'How beneficial would the chatbot be for learning about CU procedures, products and services?' saw a similar increase from 50% to 100% from week 1 to week 4. Maxim attributes this positive trajectory to his team being able to easily make changes and improvements to the virtual agent in real-time, adding more relevant content as time went on.

MSUFCU also tested the virtual agent's capacity for onboarding new staff. Employees that had recently joined the company were able to ask the virtual agent first, when they had questions about procedures or policies, before turning to colleagues for help. The chatbot's usefulness was scored at 75% when it came to helping new employees get acclimated to their roles.

"The greatest feature of a chatbot is that people ask it whatever they are thinking - you don't have to guess," added Maxim. "Boost.ai's user-friendly software makes it easy to analyze conversation data and use it to make the chatbot more functional. We watched usage and accuracy improve as more and more interactions happened."

By the end of the month-long pilot, it was calculated that MSUFCU could use the virtual agent to automate approximately 2,000 employee-to-employee interactions each month, saving time and greatly increasing operational efficiency.

An agile partnership that demonstrated real ROI

Part of the advantage of using boost.ai's no-code conversational AI platform is that it allowed MSUFCU to seamlessly integrate its virtual agent with Sharepoint, the credit union's existing knowledge base system. Employees were able to query the chatbot for what they needed and preview the relevant documentation (along with a link) within the chat panel without needing to interact with Sharepoint directly. Maxim said that employees responded positively to this functionality as it saved them a lot of time and frustration.

The partnership with boost.ai to develop MSUFCU's virtual agent falls under Maxim's purview as part of an initiative called 'The Lab at MSUFCU' - an innovation incubator that focuses on cutting-edge fintech projects including customer service automation tools. The program allows the credit union the flexibility to be agile in trialing solutions that benefit its members without needing to commit extensive resources upfront.

"It was incredibly important for us to be able to try boost.ai's solution before committing to the project and that is precisely what The Lab at MSUFCU is designed to allow for," said Maxim. "Because we were able to build and deploy the project so quickly, it gave us the fuel we needed to take to the various stakeholders at the company. We could share the experience and feedback from employees and demonstrate a real ROI by the time the process was over."