How KB Financial Group navigates and curates banking

At KB Financial Group, Dong Whan Han helps make banking better by relaying technical finance topics simply and encouraging his team to be learners and instructors simultaneously. In this conversation with McKinsey’s Eunjo Chon, Han—the company’s senior executive vice president—shares his banking and leadership strategies, the importance of a learning mindset on a teaching approach, and how to balance AI with traditional ways of working. An edited version of their conversation follows.

Eunjo Chon: If you were to describe KB Financial Group as an AI bank, what would you say?

Dong Whan Han: If you go to any KB branch, you will see a guiding kiosk that has an AI virtual human. They point the customer to the teller they should go to. As of 2022, a virtual assistant is also available inside our mobile banking app to provide guidance. We also created a chatbot that uses natural language processing to understand the Korean language and is specialized in financial terminology to complete general tasks.

We also developed the KB STA, which stands for “state of the art.” Based on state-of-the-art text analytics, customer requests and queries can be naturally generated into language, voice, or text. While we’ve made a lot of progress, we have not yet reached the engagement level—just the kiosk level—so we still have work to do.

Eunjo Chon: Many people say AI and digital-innovation companies are too focused on filling technological capability gaps and gathering data, so they become detached from reality. What is your view on this?

Dong Whan Han: That was a big problem for us as well. We don’t want to pass inconvenience to the customers by promising them the future first. AI can’t do everything, and you can’t do banking just because you have the technology. For example, if the AI service is at 20 percent, we want to fill in the remaining 80 percent with personal interaction so 100 percent of the service satisfies the customer. Otherwise, if the technology advances and AI can do 90 percent, then we’d fill in the remaining 10 percent. So it’s important for those on the business side to collaborate well with data experts and AI experts. As of now, it is very difficult for data or AI experts to have a relationship with business experts.

We want to acquire many customers, but in the digital era, customers want to receive professional services. However, our bank is currently in need of experts to provide those kinds of services. So to cover all that customer demand, we use machine learning to gather data from experts’ work and replicate it for future uses. Those in our CoE [center of excellence] support the business from the back. For this to work properly and create more value, people in the business who know the business domain need to know about the AI and data side through the CoE. Once this is achieved, we can provide more satisfactory service. We should avoid thinking that everything will work out once the technology is adopted.

Eunjo Chon: In your opinion, what level of AI should KB aspire to achieve?

Dong Whan Han: I like to think about the optimal level of AI using the concepts of navigation and curation. I often compare it to modern abstract art. If you look at abstract art, it’s difficult to distinguish the price and value. The person who can help you with that is the curator. Maybe a customer goes to the mobile app to get a loan or make a deposit because the teller guided them there. The bank’s target for AI service, then, should be to provide the curation of the product to properly sell it.

Eunjo Chon: That relates to your earlier point about filling in the 80 percent. When the bank provided only face-to-face services in the past, the bank employees were the ones providing the curation. How do you think KB’s approach to product curation is differentiated compared with other competing financial institutions in the world?

Dong Whan Han: Many of our competitor banks focused on chatbots while our AI team focused on AI OCR, which stands for “optical character reader.” Much of the knowledge in the bank is in text format. Converting the accumulated knowledge in the bank and structuring unstructured data is critical. Our AI OCR reads this data and learns from it. What kind of impact did this have? For example, the most difficult thing for me when I did foreign exchange–related work at the bank was reviewing letters of credit. It was a simple task, but it was difficult to determine what the provided information meant and what I had to do. Once we started using OCR, all that information was instantaneously read, and this resolved the pain point experienced by so many employees.

This also allowed our business side, which had only a vague idea of what AI services are, to have a much clearer understanding of how to use AI. So because KB first pursued OCR rather than a finance assistant, the large knowledge base of our bank was learned quickly and data was used more effectively. We generated outputs such as a more simplified KYC [know-your-customer] process and reduced complexity in foreign exchange work, remittance, and more. By achieving such small wins, we gained a deeper perspective on AI and we obtained more insights, allowing us to challenge ourselves to reach the next level.

Eunjo Chon: Going forward, in which areas do you think you will invest the most?

Dong Whan Han: I believe the fundamental nature of the banking business is to field customer questions and provide the answers. AI must understand the natural language and answer questions well, so the model must be continuously updated. Enhancing the capabilities to link language-learning technology to the core knowledge of our bank will be important, so we are focusing on that.

Software will keep getting better, so we can’t get tied up with a particular software and get behind on the evolution. When open-source technologies related to AI are developed externally, we should be able to create tools that can advance at the same pace, rather than buying it all at once in a package. It’s critical to be prepared for continuous updates.

Eunjo Chon: What has been the greatest success based on what you have implemented so far?

Dong Whan Han: The biggest success was KB’s AI OCR. Our AI team has also developed a knowledge graph, which distinguishes the taxonomy of certain finance-related words—volatility, for example—and updates our knowledge base accordingly. We also consolidated our call centers, which is a challenging task. In finance, the banking business, insurance business, and securities business have different characteristics, so to consolidate all of them into one call center is not an easy task. But it is feasible if we leverage AI. KB is planning to apply a consolidated AI call bot in Korea. The fact that we are doing it as a group and not just at the bank level may be our differentiator and success factor.

Eunjo Chon: If you were to give some advice to other financial institutions or companies that want to drive innovation or invest, what would it be?

Dong Whan Han: Microsoft CEO Satya Nadella said, “The learn-it-all does better than the know-it-all.” I think that is very important. We should make the effort to keep learning and trying new things with an open mind. And we should provide the framework for people to enjoy learning. For instance, KB is recruiting many AI or digital experts from outside. But I’ve noticed that sometimes we use them as translators. AI has a lot of technical elements, so many people on the business side don’t try to understand it. But if you remove the frills and just focus on its essence, it becomes quite easy to understand. You shouldn’t pretend to know something when you don’t—remove the noise and focus on the core message.

Eunjo Chon: You have worked at KB for a long time, serving in many key positions. How has your role at KB changed over time?

Dong Whan Han: Honestly, I joined the bank to make a living after getting married. Over time, I realized that while the government’s role is protecting the lives of its people, the banking system’s role is to protect people’s economic happiness. The bank where I was working helped me to make a living during the financial crisis, which is quite meaningful, and I realized that banking played a big role in protecting the happiness of people around us, too.

While the government’s role is protecting the lives of its people, the banking system’s role is to protect people’s economic happiness.

Dong Whan Han

Over time, I became comfortable sharing my opinions and made proposals for different solutions. That created a good impression and allowed me to work on the strategy part of the business. So I emphasize to my colleagues that they should also find what engages and interests them in their work and work for that thing, rather than for their company or boss.

Eunjo Chon: How can leaders help employees be successful?

Dong Whan Han: I have talked frequently about being a “top-notch instructor” to my staff or during executive meetings. Top-notch instructors refer to the lecturers in Korea who prepare students for college entrance exams; they often have great skills in delivering knowledge about, for example, math—better than schools do. Finance is abstract and very difficult to understand. So we should benchmark train our financial experts to be top-notch instructors and help them deliver financial content better to our customers. When customers are just told to sign a contract without knowing the exact content, it could lead to mis-selling, which would trigger issues and complaints when losses occur. So how financial companies deliver accumulated expertise and knowledge effectively to customers matters for their success.

Eunjo Chon: What will it take to nurture these top-notch instructors?

Dong Whan Han: When you look at our products, they are represented in documents such as manuals. But today, it’s all about videos. Experts, however, find it difficult to give explanations on a video. They are not trained to explain a product’s specialized features in layman’s terms. Fintechs in Korea, especially KakaoBank and Toss, have recently made good progress, albeit in small areas, thanks to their ability to explain their products simply. So for our wide business areas, including insurance, securities, and banking, the key would be to offer good explanations by tailoring them to our customers.

By taking the “learn it” approach and communicating easily with our customers, we can allow our top-notch instructors to deliver our hard-earned financial knowledge. We already have accumulated knowledge, but how to deliver it well has been less of a focus. To help our customers understand information better and faster, special training would be required.

Eunjo Chon: What achievements are you most proud of? And are there any areas that you want to improve?

Dong Whan Han: First, I am proud of our people. While before we had strong knowledge in domains, our knowledge in digital technology was insufficient. Even so, we had to provide services and compete, so we worked together. The people whom I worked with at that time have developed themselves and are now serving in key positions across our organization. Second, KB Star Banking, which is the mobile platform of our bank and the group, has grown by more than 14 percent this year alone, achieving more than ten million MAU [monthly active users]. Fintech companies have achieved similar MAU by operating in very narrow areas, but our services are wider and more complex.

As for the areas in which I want to see improvement, it’s about Big Tech. Those firms are our rivals, so we often try to follow their practices despite them being complicated and not always best. We should think instead about simplifying our banking platform as needed. That would also help assure that our ten million MAU are really satisfied. We look forward to accomplishing those goals.

Explore a career with us