Finn AI Customers share practical lessons learned from their conversational banking journeys.
While in the past they’ve seemed to be available only to national banks, AI-Powered Chatbots are becoming more feasible to implement for regional and community banks and credit unions as well. Conversational banking solutions delivered as a SaaS allow for banks and credit unions to implement a virtual assistant while mitigating much of the time, resources, and learning that would normally be required for such a project. Now, more banks and credit unions than ever are able to bring a modern digital customer experience to their own digital and mobile channels.
BECU, Allied First Bank, and Civista Bank are three examples of financial institutions that have recently taken the plunge into exploring AI-powered chat for their digital customer service experience. Their experiences so far have all been successful, but as this was new territory for them all three experienced unknowns they hadn’t considered or initially addressed in their project plans. At the Conversational Banking Summit, leaders of these institutions came forward to share their own experiences implementing a chatbot, and each of their stories shares with us a different lesson from their journey.
BECU: Think Through Your Limits
Back in 2017, BECU was noticing a particular pain-point with their customer service: they couldn’t operate their call centers or message inbox 24/7, and subsequently members weren’t always able to have their urgent banking needs met when they truly needed them. Having a way to serve these members outside of their usual business hours, and ideally with a method that would bypass a live agent altogether, was the ultimate goal BECU hoped to achieve. So, after a successful run with instant-messaging based customer service, BECU decided to implement an automated chatbot of their own.
BECU had no prior experience designing or running a chatbot, so they started very small with a simple bot that could answer rudimentary questions for customers at any hour. It worked well at first, but as member’s expectations for the chatbot grew beyond what BECU was capable of providing. As the bot began to become unable to answer user questions, they knew they would need to outsource further development to a team with more experience with conversational banking.
“As we added more complex or longer content, we could see that it started to test the limit of what a member was willing to do with the bot,” said Amie Bosshart, Digital Channel Manager at BECU. “As the content expanded, and the intent categorization grew, we also saw a dip in accuracy”. BECU then knew they needed to improve the bot in two ways: growing its natural language understanding, and helping users get a seamless way out to a human when they needed it. This would take them a lot of time and effort, however, and Amie continues that “If we wanted to grow our AI offering, we needed to be more agile and more scalable”.
By choosing Finn AI to help expand their chatbot’s capabilities past what their original bot solution was capable of performing, BECU could leverage their members’ pre-existing trust with the bot and create an even better experience while trusting the development to experts in the field. They knew not to rush too quickly into the project, prioritized quality over quantity, and turned to specialists when they reached their own limits of knowledge on AI-powered chatbot development.
Allied First Bank: Optimize the Human Element
Allied First Bank was experiencing a slightly different issue: a high volume of low-importance customer queries. A large percentage of the calls to their contact center were for very basic, straightforward navigation questions or services. They felt it was distracting their support staff from tending to more urgent needs that require human intervention due to their complexity.
“It comes back to the high volume, low value questions and requests customers have when they come online and come to your website,” said Rory Bolen, Director of Marketing at Allied First Bank. “People need the actual more in-depth, person-to-person, contact response and solutions”.
With the pandemic causing a general increase of people seeking customer service at-home, this pushed the issue to a breaking point. An AI-Powered chatbot was their answer: able to automatically resolve simpler user goals and routing them to human support staff when needed.
Somewhat ironically, the implementation of their chatbot was actually to allow for more personal, human interactions with customers. Adding their chatbot wasn’t to prevent customers from speaking with their live agents, but simply to free up their time to allow for much more intimate and helpful conversations. There are many things that even the most intelligent robot can’t do, and improving their ability to provide real human service was an important part of the chatbot implementation for Allied First. While 80% of customer inquiries could be solved with a virtual assistant chatbot, for the remaining 20% the chatbot can easily send a user to a live human agent for more complex, personal interactions.
Civista Bank: Avoid Multitasking
Pre-pandemic, Civista Bank was in the process of a 5-year plan to help develop their digital customer-facing channels. Once the pandemic happened, Civista rapidly increased the pace at which this plan was being put into action: providing at-home service for their customers was now an imperative. Through their existing partner AnyHour Solutions, Civista learned about Finn AI and how AI-Powered Chat could seamlessly achieve this goal of theirs. Suddenly, getting a banking chatbot for their online platform seemed like a no-brainer.
“We began conversations with both Finn and another vendor to talk about what that could change for us,” said Donna Jaskolski, SVP Customer Experience Officer at Civista Bank. Adding AI-Powered Chat to their online bank was very much an exciting prospect for Civista, however, as they were currently in the middle of their digital transformation, they soon realized not to rush into this new prospect. “Really thinking strategically about what that looks like, we began having conversations with Finn and realized very quickly that we would kind of have to backburner what that conversational chatbot would look like for us in the future.”
What Civista learned throughout all this is to maintain priorities on projects: deciding on a chatbot in the middle of their digital transformation proved to be a difficult task, and careful management of their ongoing tasks was needed. They knew that an AI-Powered chatbot would be no small task, and it needed to have all of their focus before they could really dedicate themselves to the project. “Understand that your marketing team and your product team are going to have a lot of heavy lifting for about 30 to 60 days on implementing your chatbot” Jaskolski continues, “because it’s really on them, and the entire organization, to put a voice and personality onto it”.
There’s no single chatbot journey for everyone: each of these organizations had their own unique desired outcomes for their conversational AI solution. All of them did, however, have a unique experience which can impart lessons on those looking to try conversational AI themselves. All three journeys have been successful, improving customer satisfaction and enhancing the overall banking experience. Virtual assistants are a new area to explore for many, however, so observing how the chatbot journey goes for others can give a financial institution key insights and make their own implementation even smoother. As more banks and credit unions begin to implement AI-Driven chatbots, it only becomes easier for financial institutions who have yet to explore conversational AI to take concrete lessons from these peer innovators and incorporate them into their own digital strategies.
You can watch the full Conversational Banking Summit session recording here