By partnering with Observe.AI, the leading regional lender is improving call categorization, boosting agent performance, and enhancing overall customer experience.
With deep roots in community banking, Central Bank serves customers across eight states, mainly in the Midwest. The $20-billion privately-held bank supports local decision-making and delivers financial solutions through a network of more than 150 regional banking facilities.
At the core of the company’s service model is its Customer Service Center (CSC). It handles more than 3,000 customer interactions daily through phone, email, chat, and mobile apps, ensuring personalized and locally connected service.
“We’re a group of community banks,” says Mary Beth Gillum, Sr. Vice President of Customer Service Center. “It’s very important to us to provide ‘legendary service’ while maintaining that local look and feel.”
However, reliance on manual and time-consuming quality assurance (QA) processes made it difficult for Central Bank’s contact center operations to gain the insights needed to improve customer service. Basic customer relationship management tools also limited the company’s ability to effectively track and analyze customer interactions.
“When I started, we used big metal binders—what I call ‘parts books’—to use as a resource for everything,” recalls Gillum. “Representatives would take a call, choose a topic from a list, and click a button to track the call reason.”
To address these inefficiencies, Central Bank turned to Observe.AI.
“What swayed us was Observe.AI’s user-friendly interface,” says Jeffrey DeBourge, Contact Center Manager and AI Administrator at Central Bank. “Whether it’s an intern in college or a senior analyst, everyone can use it with no issues.”
– Jeffrey DeBourge, Contact Center Manager and AI Administrator, Central Bank
Rolling out Observe.AI’s Post-Interaction AI has addressed several pain points for Central Bank. By automating QA processes with features like Auto QA, Screen Recording, Moments, and searchable transcripts, the company has transformed its contact center operations.
Automating the analysis of all customer calls has enabled the CSC team to gain deeper insights into customer interactions, which in turn helps them analyze sentiment and increase agent effectiveness. In the third quarter of 2024, the team evaluated 167,000 calls, a jump from just eight per month or 24 per quarter before adopting Auto QA.
“The QA portion was the first thing we focused on,” says Gillum. “The two other key components were automatically capturing the information from a call and making it reportable, and sentiment analysis or understanding what people are calling about.”
DeBourge adds that his team has seen tremendous value with Auto QA. “We’ve also had a lot of success identifying the reasons why our customers are calling. It’s been a really valuable process, and we’re very happy with the results.”
The benefits of Post-Interaction AI quickly became clear, as the team could quickly access and examine all available data pertaining to customer interactions across different channels. This has enabled the team to gain actionable insights into customer needs, address common issues more effectively, and improve overall service delivery.
“Before, it would have taken hours to investigate a customer’s issue. Now, we can search keywords and pull up relevant calls in minutes,” says DeBourge.
Agreeing, Gillum notes how the Observe.AI platform has accelerated the monitoring and reporting of call trends. “During a global IT outage, I searched keywords like ‘global IT outage’ and a couple of other related phrases. Within two minutes, I identified calls related to the outage and gave the leadership an accurate count [of customer issues],” she adds.
One of the most notable changes for the CSC team was eliminating disposition codes, which are manual labels agents use to categorize calls. By replacing these with AI-automated tracking using Moments to identify a more granular set of key events and themes taking place within calls, the team has streamlined issue categorization, improving both efficiency and accuracy.
“The change reduced our after-call work to about 10 seconds from an average of 30 seconds previously, which is exceptionally low for a financial call center,” says DeBourge.
– Jeffrey DeBourge, Contact Center Manager and AI Administrator, Central Bank
Importantly, the CSC team has also identified and addressed agent behaviors that unnecessarily increase time spent on calls.
“We had an agent who was staying on the line to make sure the transfer was successful, but it was increasing her handling time,” shares DeBourge. “We uncovered this through Auto QA, and she immediately adjusted her behavior after seeing the data.”
Likewise, supervisors can now spot silences and hold time behaviors, and visually observe an agent’s follow-through activities with the Screen Recording feature. “We’ve started focusing on outliers,” says DeBourge. “Some agents don’t put customers on hold, while others use hold for everything. Observe.AI helps us find the right balance and rethink our strategies.”
Identifying key agent behaviors has helped the CSC team develop data-driven plans and reduce average handle time by up to 5%, resulting in significant efficiency gains.
“Our call center teams handle between 2,000 and 2,400 calls every day,” says Gillum. “If your staff reduces their after-call work time from 30 seconds to 10 seconds, the time savings can be equivalent to three staff members.”
By detecting behaviors that affect performance, such as excessive hold times, the team has also refined its pay structures for agents to reward efficiency and quality.
– Mary Beth Gillum, Sr. Vice President of Customer Service Center, Central Bank
Post-Interaction AI’s strengths extend to improving customer interactions and detecting fraud. By harnessing its ability to analyze call drivers, the CSC team now better understands why customers are reaching out.
“Fraud is a prevalent issue,” explains DeBourge. “We’ve built Moments to identify fraud-related calls, and we’re exploring deeper insights to understand their root causes.” Using analytics and insights from Moments, the bank is developing self-service tools to help customers manage fraudulent charges more effectively and proactively.
The CSC team has also created Moments to identify service opportunities during calls. For example, when a customer mentions debit card fraud, agents are prompted to recommend setting up online account alerts to boost security.
This practice has provided useful training examples to promote consistent standards among agents. Previously, sales managers would listen to multiple calls to identify effective techniques. “Now, they can search for high-performing phrases used by top agents and incorporate them into training,” says Gillum. “It’s a huge time-saver.”
Central Bank is just beginning to explore the full potential of Observe.AI. With plans to expand the analysis of customers’ reasons for calling and to strengthen fraud detection, the bank aims to gain deeper insights and deliver better customer experiences.
“We’re building out more advanced insights to understand the nuances behind common call drivers like fraud,” shares DeBourge. “That will help us not only address root causes but also improve our communication with customers.”
The bank also plans to roll out Observe.AI to other areas of the organization. For example, its resolution center—an internal help desk for bank procedures—is looking to use transcription and analytics to identify call drivers and enhance employee training.
“It’s exciting to see other business units adopting Auto QA for coaching and increasing team performance,” remarks DeBourge. “Everyone in our company wants Observe.AI.”
Central Bank is a privately owned financial institution headquartered in Jefferson City, Missouri. With $20 billion in assets, the bank provides community-focused banking services through a network of more than 150 locations and access to over 22,000 ATMs across the United States. Central Bank’s operations span Missouri, Kansas, Illinois, Colorado, Iowa, Tennessee, North Carolina, Florida, and Oklahoma.
.jpg)
.png)
.png)
.png)
