Covering the banking industry, I find myself using artificial intelligence (AI) more than ever. Recently, AI reportedly outperformed human experts in commercial banks' wealth management services, and after trying it myself, I understood why. Explanations grounded in market and customer data seemed to inspire more confidence than those from a human advisor.
But the assessment in customer service is entirely different. Many users say the AI chatbots embedded in banking applications simply fail to hold a conversation. The same goes for phone consultations. Search "AI consultant" on portal sites, and you will find countless reviews from customers who waited forever for a human agent, never got connected, and eventually gave up. Why is the same AI so trustworthy on one side and so frustrating on the other?
The difference lies not in performance but in where it is being used. AI shows its strengths in wealth management, where it produces results based on data, but its limitations surface in consultations, where it must handle vague customer language and unexpected situations. Because it operates around scenarios customers frequently ask about, it fails to properly answer specific questions such as, "With an annual salary of 60 million won and a credit score of 870, how much can I borrow?"
Still, banks are pushing AI into the consultation space. The first areas to shrink are call centers and branches. The number of consultants at the four major banks, which stood at 3,627 in 2022, has dropped by more than 400 in three years. Over the same period, branches have decreased by 195, making workforce reductions a reality.
For banks, this is management efficiency; for customers, it is an inconvenience. Helping customers who hesitate in front of a screen, unsure of which button to press, or patiently listening to elderly customers unfamiliar with financial jargon, is work that cannot be entrusted to AI. It is a human's role. "Customers who come to the branch are here to ask staff directly, no matter how long the wait," an employee at a commercial bank branch said. "If staff numbers drop further, the customers will ultimately bear the loss."
The difference between good AI and bad AI does not lie in technology alone. Where and why it is used matters more. Banks' AI must become a technology that helps customers, not a tool to cut staff and save costs. If AI exists only to boost profits, no matter how advanced the technology becomes, trust will not follow.
