An office worker in her 30s, a person surnamed Kim, recently tried using an artificial intelligence (AI) chatbot to learn about an indemnity health insurance product from DB Insurance, only to give a wry laugh. She typed "please introduce an indemnity insurance product" into the chat window, but the reply came back: "This is something I haven't learned yet." Kim was able to confirm the information she wanted only after speaking directly with an employee.
Although the financial sector has been increasingly introducing "AI agents" recently, many customers say they are "frustrated." AI often fails to provide the desired answers, and many companies require customers to go through the AI agent before they can speak with an employee.
According to data released in Jul. 2024 by market research firm Reports Globe, the size of the AI market in Korea's insurance industry is expected to grow from about $150 million (about 200 billion won) in 2022 to more than $550 million by 2028. Growth potential was found to be particularly high in areas such as customer service and fraud detection.
Domestic insurers are also rapidly expanding chatbot and AI consultation services. Meritz Fire & Marine Insurance has introduced a customized AI consultation assistant with recommendation and analysis features, the first of its kind in the non-life sector, and DB Insurance has launched the AI chatbot "Promy." KB Insurance has also introduced an AI-based consultation agent to support customer service.
However, actual consultation services still show limits. AI responds efficiently to simple questions but often fails to handle complex product inquiries. When specific products such as "insurance for people with preexisting conditions—cancer" or "children's insurance" are entered into the chatbots operated by Meritz Fire & Marine Insurance, DB Insurance, and KB Insurance, in most cases the bots fail to answer or connect the user to a human specialist.
These limits are cited as a factor that lowers customer satisfaction. According to the report "Financial consultation innovation through AI–human collaboration," released by KB Management Research Institute in Oct. last year, only 21.6% of adult men and women in Korea surveyed said they were satisfied with AI consultation services at financial company call centers. Among those dissatisfied, 73.6% said AI, such as chatbots or voice recognition, did not properly understand their requests.
The side effects also add to the burden at consultation sites. Inquiries that AI cannot resolve are connected to human agents, but by then customer dissatisfaction has grown after time was already taken by AI. According to another survey cited in the same report, 66% of insurance consultants said their working conditions had worsened since AI was introduced. The most common reason, at 53%, was "increased customer complaints due to AI errors, which raised stress."
An insurance industry official said, "For now, the lack of expertise among AI agents is causing inconvenience for customers, but as more data accumulates, consultation quality will gradually improve," adding, "Ultimately, it will develop in a direction that increases the efficiency of consultation work."