"Customers Are Already Angry... We Need to Change the Way AI Is Applied" [Bbaengbbaengi AI Call Center]⑥

Meeting Three 8-10 Year Veteran Call Center Agents in Daejeon, the 'Call Center Mecca'
"Elderly Customers Avoiding AI Consultations Call Branches... AI Confusion Everywhere"
"Need to Change Application Method... AI Should Be Redesigned to Support Agent Consultations"

Editor's Note“It’s a ‘person,’ not AI, dear customer. How can I assist you?”
Call centers, which everyone has used at least once, have transformed from places where agents kindly answered calls and resolved issues into inconvenient spots for financial consumers. With the emergence of the unfinished ‘Artificial Intelligence (AI) consultation service,’ the problem-solving process and time perceived by financial consumers have only been delayed. This irony has arisen as banks, card companies, and other financial institutions replace human agents in call centers with AI consultation services. Accordingly, Asia Economy aims to look into the inside story of financial company call centers from various perspectives, including financial consumers and workers.

“Nowadays, after customers get frustrated from going around in circles with AI chatbots and voice bots, we have to be careful even about breathing.”

Hyun Jin-ah (43, female), Ban Soon-geum (50, female), and Kwon Young-woo (49), call center agents of financial companies whom Asia Economy recently met at Vision Square in Seo-gu, Daejeon, shared the difficulties they have experienced on-site since the introduction of AI consultation services. These individuals are ‘veteran agents’ with 8 to 10 years of work experience.

In financial company call centers, contrary to expectations, the workload of human agents has rather increased since AI consultation services became widespread. Although the plan was to handle simple inquiries with AI, in reality, the increased procedures and heightened emotions have resulted in accumulated dissatisfaction and fatigue for both customers and workers.

While there is consensus that the advancement and expansion of AI consultation services are an inevitable trend of the times, there are calls to improve how these technologies are applied in work processes, given that the technology is still incomplete and has clear limitations.

Daejeon Kookmin Bank Call Center exterior. Photo by the Solid Call Center Branch of the Korean Public Service and Transport Workers' Union, Korean Confederation of Trade Unions

Customer Service Starts with AI... Delays Lead Customers to ‘Lose Patience’ and Agents to ‘Be Flustered’

The most practical problem agents face is the relationship with customers. Financial companies encourage connection to AI consultation services such as chatbots and voice bots when using call centers, which means customers have to go through additional procedures to resolve their issues. Hyun, an 8-year veteran call center agent at Hana Bank, said, “These days, AI first answers the call, and if AI cannot understand or handle the task, it connects the customer to an agent.”

In the case of domestic card companies, even if customers call the customer center, the structure does not allow immediate connection to a human agent. For example, customers must go through automated response systems (ARS), AI chatbots, or AI agents (voice bots) before they can be connected. Bank call centers also mostly require customers to first talk to an AI agent and then connect to a human agent if desired.

Because of this, customers unfamiliar with the system, including middle-aged and older adults, often experience significant delays in connecting to agents or fail to have their requests processed in a timely manner. As it becomes harder to connect to human agents, there are even humorous posts on social networking services (SNS) asking how to reach call center agents. This situation negatively affects both agents and customers. Hyun said, “Elderly customers want to talk directly to agents but complain a lot about why they can’t get connected,” adding, “Because of this, many people call the general branch numbers expecting to be connected directly to staff.” However, to talk to branch staff via general branch phone numbers, one must know the internal extension number.

Especially, many calls to call centers are simple inquiries like account checks that could be resolved quickly, but AI often fails to understand the customer’s inquiry, causing longer processing times. Ban, who has worked as a call center agent at KB Kookmin Bank for over 10 years, said, “Customers complain that AI doesn’t understand what they say and just makes them go around in circles,” and added, “Customer attitudes have worsened compared to before AI consultation services.”

A KB Kookmin Bank call center counselor at work. Photo by the Solid Call Center Branch of the Korean Confederation of Trade Unions, National Public Transport and Social Service Workers' Union

‘Emotions’ AI Cannot Replace... Sometimes Leading to Angry Calls

According to the agents met that day, many problems arise because AI agents find it difficult to fully understand customers’ demands and emotions. Complaints that could be resolved through emotional communication between humans often turn into ‘angry calls’ when handled by AI agents.

Kwon, a 10-year KB Kookmin Card agent, said, “Even if AI perfectly handles simple tasks, the problem is whether it can understand and judge the emotions behind customer requests.” Hyun added, “There are many tasks that could be resolved faster if the customer’s feelings were comforted, but AI agents cannot do that, so they sometimes hinder the process.”

Agents firmly stated that even with AI advancement, there are unique roles that only humans can fulfill. Tasks like debt collection are difficult to replace with AI. In fact, Kwon shared an experience where he proposed a long-term card loan alternative to a customer suffering from snowballing revolving interest, reducing the monthly payment and helping the customer secure living expenses. Such work is beyond the scope of AI, which can only respond to customer questions. He said, “Even if AI stabilizes, professional consultation is necessary,” and added, “Agents assess customers’ situations, suggest alternatives, and provide advice.”

Of course, the human agents on-site acknowledge that the advancement and expansion of AI are inevitable trends. However, they unanimously agree that urgent measures are needed to improve call center work efficiency in ways that do not cause inconvenience to financial consumers as currently experienced.

For example, some financial company call centers currently apply AI that transcribes conversations between agents and customers, but even this often fails to function properly because it cannot accurately recognize terminology. Also, although there are bots assisting consultations, they are not advanced enough, limiting their usefulness. Ban said, “AI is no help to agents. We still have to look up product lists one by one in the table of contents during consultations,” and added, “AI should develop in a way that helps agents provide better consultations.”

There are also calls for urgent alternatives for vulnerable groups such as the elderly. Hyun emphasized, “For elderly customers, explanations need to be detailed down to ‘what to press at the top right of the homepage,’ but AI alone has limitations,” and stressed, “As society ages, the need for human agents will increase.”