By W.B. King
With the continued goal of making banking easier and more convenient for its 106,000-plus members, Unitus Credit Union undertook a three-year strategic plan in 2020 that included offering an enhanced virtual branch. COVID-19 accelerated the process.
“We had planned on rolling out virtual teller machines followed by chat and video banking. Once the pandemic hit, in an effort to keep both members and employees safe, branches closed lobbies and we switched to drive-thru models only,” Jessica Smith, AVP of Remote Experience at Unitus Community Credit Union, told Finopotamus.
Due to shutdowns and safety protocols, Smith explained that the $1.7 billion, Portland, Ore.-based credit union’s call volume increased by 38% in April 2020. “This led us to switch the plan to partner with Glia to launch chat first in April of 2020 and then we immediately followed with launching video banking in July of 2020.”
Unitus Community Credit Union is among approximately 400 credit unions, banks and insurance companies the New York City-based Glia counts as clients. The Glia Interaction Platform solution aims to unify digital customer service (DCS), traditional call center, and automation with its ChannelLess solution.
Within the first month of launching Glia Virtual Assistant (GVA), Smith said the credit union realized “impressive results.” GVA, she explained, handled nearly 1,100 engagements with 52% of those interactions ending without being transferred to a human.
“Over half our members’ questions are being answered 24 hours a day, seven days a week without needing any human assistance. During the first month our GVA also had an understanding level of 89% and a member satisfaction score of 4.2 out of a 5-point scale,” she continued. “We quickly realized this was a great addition to our commitment of meeting our members where they are, when they like.”
While she noted that the credit union doesn’t track at the account level to determine percentage of membership using the solution, it does track at the transaction count level. “Nearly 60% of our remote channel transactions each month are initiated with the chatbot,” she said, adding that other remote channel transaction types include email, audio and video.
Hablas Espanol? Si
Based on the success of Glia’s solution along with lessons learned, Unitus Community Credit Union’s Senior Vice President and Chief Retail Officer Corlinda Wooden explained that a potential coverage gap was identified.
“As part of our commitment to providing inclusive financial services in the communities we serve, we knew there was a need for a Spanish-speaking chatbot,” she told Finopotamus. “At Unitus, we have seen Hispanic member growth reaching 32% within four years.”
Smith further explained that within a three-mile radius of each of its 12 branches, over 116,000 people speak both English and Spanish. She added that 47 of the credit union’s 330 employees are Spanish-speaking.
“We want to be able to reach our members and community where they are in their preferred language,” Wooden told Finopotamus. “Unitus Community Credit Union is dedicated to reaching this emerging market by offering communications in both English and Spanish.”
When the new solution rolls out on June 30, 2023, Unitus Community Credit Union will be the first credit union in the Pacific Northwest to launch a Spanish chatbot, Unitus Community Credit Union’s CEO Steven Stapp proudly stated.
“Nearly 9% of Oregonians speak Spanish as their primary language in the home,” said Stapp. “We believe communication and relationship-building is at the heart of our service model, and this new tool will help improve access to anyone who prefers a Spanish-language interaction.”
Glia’s Virtual Assistant Specialist Jake Tyler echoed Stapp’s comments.
“Unitus has a large and growing Hispanic member base and this further underscores their deep commitment to meeting these members on their terms and in the language of their choice to make banking easier and more convenient to use,” he said.
The Tech Bundle
Smith noted that Unitus Community Credit Union utilizes the “Glia Hub,” which includes chat, audio and video engagements, live co-browsing, and English and Spanish bots. On the Call Center side, the credit union uses Glia Call Visualizer Chat and Co-Browsing paired with the phone system.
“Glia Hub allows the Virtual Branch team to meet face-to-face with video as well as answer live chats with the option to upgrade to audio or video,” Smith continued. “Our English and Spanish Virtual Assistants is available for 24/7 and Glia Call Visualizer and Co-Browsing allows our employees to collaboratively guide our members on-screen, ensuring we are walking with our members from start to finish.”
To rollout the solution, a credit union tech staffer was needed for “configuration set up” in the Glia Hub for staging and production, Smith said.
“The heaviest lift was in the initial creation of the content that was performed by the Virtual Branch team. Glia provides responses (intents) for the Virtual Assistant to answer our members’ questions,” she noted. “Those intents need to be personalized to fit our brand, products and services — 380 intents are required to launch. This process required 40 to 60 hours of work with an additional 20 hours of review and another 40 to 60 hours of testing.”
Each GVA component rolls out separately, Smith explained. “We have a public desktop GVA live and a desktop authenticated GVA (within digital banking) live. Next will be our Spanish public desktop GVA followed by our Spanish authenticated GVA. We will then mirror that process with mobile devices.”
The first GVA, Smith added, was the “heaviest lift” with a timeline from contract signing to full roll out about three months. Timing can vary based on “various business decisions along the way and resource availability,” she offered.
“Glia estimated a six-week project timeline for the first GVA. The process for our second GVA took about 10 weeks to complete and was simplified because we were able to clone some of the content and had become proficient with the content writing tool and process,” she continued. “We anticipate the time to launch our additional GVA will continue to decrease. Especially without mobile versions since the content will be the same.”
Along with an approximate 50 hours of testing in staging, the credit union also performed what Smith referred to as a smoke test.
“We had the Virtual Assistant turned on with our lowest traffic page to test functionality in a live setting prior to rolling out on all pages,” she said. “Again, these time frames could vary based on business decisions and resources. We were very intentional with our testing to ensure the accuracy and value-add for our membership.”
Chatbot Persona
Since the launch of both Glia Virtual Assistants, on the credit union’s public webpage and within its Digital Banking platform, there has been an 85% increase in chats answered and a 63% increase in average speed of answer.
“We have also seen a reduction of 60% in our average handle time and an additional 63% reduction in our average abandonment rate,” Smith said. “In total, our Virtual Assistant interacted with our members for 330 hours and answered 6,344 chats with 42% of those chats being brought to completion without having to be transferred to a human.”
When Finopotamus asked Wooden how these initiatives will impact both employees and members moving forward, she responded that the initiative represents a significant milestone.
“An attractive quality of the bot is that it learns so we can see what questions our members are asking and really dive into the analytics of it all to improve communication and messaging where needed,” Wooden said, adding that it “frees up our employees to do what they do best,” which is “be human, and handle those more complex issues for our members.” For employees, she said “it’s like gaining an extremely proficient member of the team.”
Since the goal is to match member communication with the ethos of the organization, Smith said credit unions looking at a similar offering should schedule the necessary amount of time for content writing. She also said the first bot will be the most challenging, noting that it’s important to determine whether or not the bot has a “persona” before content writing begins.
“Depending on your decision, that could impact the content and process. For example, if you want the bot to be friendly and fun versus professional and formal, the content is unique for each of these approaches," Smith said. “Lastly, start thinking about post-launch content management processes and ownership. Determine who owns this process and is responsible for updating and creating content post launch on an ongoing basis.”