Hello, I am Ha Hwa, the head of Hwagum Financial Group in charge of this presentation. I didn't realize the pressure of giving this presentation would be so great after the presentation by KB Woori Osangmu. I apologize. I joined Hana Bank in 1995 and have been working at the bank for 30 years now. I am a banker that you can easily encounter at a bank branch. In this dual position of discussing the future of finance brought by the age of children, I couldn't sleep well thinking about what insights I could give you, with only experience in the field of banking. But our employees said so. Those of you here would want to hear about the stories from the field. So, with confidence in that belief, I came out. Leveraging the strengths of my career, I will talk about the essence of financial services experienced in the field for 30 years and how consumers can be supported by AI in financial services. Changes, AI adoption, and my career expertise.

I will talk about cases of applying factor introduction strategies. First, I want to talk about change. First, let me ask a question and start. Please pay attention for a moment. Could those of you who have visited a bank branch in the past month raise your hand? There are five people here, and two of them were our bank employees. In the past, the competitiveness of a bank was mainly determined by the number of bank branches. Let's take a look at the screen together. The chart shows the channel usage of customers who transact with our bank. If you look at the top part of the bar, it represents customers who only transact at branches, which is now only 12%. The lower part of the bar represents customers who mostly transact digitally, which is close to 70%. Especially noteworthy are the numbers in parentheses, which represent customers who have not used a branch for transactions even once in a year, which is as high as 32%. The next screen is for Hana Bank.

I am a new customer to deal with. Now, transactions are mostly done online rather than at the branch. With the advent of online banking, the loyalty of customers I believed to be main customers while working in the field has already disappeared. It is challenging for us on business days when customers easily engage in duplicate transactions or quickly switch their main transactions to where they can profit immediately. The Korean Financial Newspaper, which organized the event, has repeatedly highlighted changes in the transaction patterns of financial consumers in its articles. In line with technological advancements and changes in consumer behavior, high-interest services are also evolving and innovating. Let's take a moment to look back at the past. If you look at the photo on the left, you can see some young people there, and that wooden board is a counting board. Even in the 1980s when I opened my first bank account using a calculator, banks used counting boards. This era is considered as Bank 1.0. Later, calculators were introduced, and financial.

We consider the period until the introduction of computerization and ATM machines as Bank 2.0. With the rapid introduction of the internet that began to spread in the late 1990s, internet banking was fully realized, leading to the start of insurance sales on home shopping internet channels, marking the era of Bank 3.0 Phase 1. Since the release of smartphones into the world, we have been experiencing the beginning of the era of smart banking until today as Bank 3.0. Now, with only curiosity, we look forward to what the next Bank 3.0 will be. Let's take a look at the screen for a moment. [MUSIC] [APPLAUSE] Have you enjoyed watching? When do you think this video was made? This video was created in 2018 by Hana Bank, contemplating the future of finance.

Some of you may have seen it before. When I watched a video seven years ago, it was like a science fiction movie. However, about half of the things mentioned there seem to be already happening. Also, because of AI, it feels like everything will be possible. Even for a non-major like me, who didn't major in AI, the word AI sticks in my mind. One of them is the movie Terminator, which I watched with great interest. You all know the story, right? When you see that content, it's about machines equipped with AI destroying humanity. However, recently, I saw news about drones used in the war in Ukraine, and yesterday, news came out that in China, they are training robots equipped with guns. It's becoming a reality. It's a bit scary, actually. And another one is the game of Go between Lee Sedol and AlphaGo.

It was a showdown. The shocking scene where a human defeated AI. I don't know much about Go, but I enthusiastically supported the human player's efforts. However, this story dates back 8 years. With the story from 8 years ago, AlphaGo, which was once ahead of us, is now once again taking the lead in the world of AI. The AI that emerged long ago is now being used in many places. In fact, we, as ordinary people, did not need to directly utilize it or consider its possibilities. But with the introduction of AI like Clova X and Reuten through the book Pete, it gave us the experience of being able to use them directly. This surprised everyone, right? I will introduce a report from the UK Finance, which is a financial institution in the UK. The association conducted a survey of UK financial companies in 23. The report shows that UK financial companies are using AI technology for analysis and generation according to their respective characteristics.

Respondents in a survey have stated that the adoption of AI is gradually increasing, especially in areas such as AI fraud detection and modeling, not only in the UK but also in Korea. It is widely used domestically. Furthermore, it is expected that AI will be specialized in the field of real-time AI business automation and customer service. According to the report, the selection and utilization of AI technology are necessary depending on the characteristics of each utilization area and situation, emphasizing the need for the creation of timely services. Just as financial institutions in the UK consider AI adoption necessary, many financial institutions in Korea are in the process of adopting AI for rapid and effective service innovation. However, there are several hurdles to AI adoption. I will share my experiences with some of the factors in adopting generative AI. As you all know, hallucination is the most significant factor. Generative AI predicts the next most likely sentence based on the learning data, creating sentences.

There are technical limitations in creating wrong answers. Providing incorrect information in the financial industry that requires 100% accuracy can lead to financial accidents. The expensive cost is also a problem. When looking at the data provided by financial security, the infrastructure investment required to directly introduce commercial models is as high as 750 billion won. The cost of investing in open models or using tokens for external models based on cloud services is also a significant burden. Furthermore, the limited use of Korean in AI learning seems to be a problem. If you look at Google's model, Korean tokens are only used 0.195% of the time. In this situation, the quality and reliability of Korean-based financial services may require more time. Security is also a major concern. Currently, the National Rights Commission has separated the internet area from the business area, but network separation poses several limitations for utilizing AI. Regardless of the benefits of AI adoption in finance, there are absolute limitations due to various factors.

There are parts that cannot be compromised. I believe that providing 100% accurate answers and financial information is essential in terms of personal information protection. Due to the nature of generative AI technology, there is a possibility of providing incorrect information, but in financial services, only 100% accurate information is required. In terms of information security, less sensitive information such as internal regulations can be utilized in external public clouds for enhanced security. However, sensitive financial personal information cannot be entrusted yet. Nevertheless, despite these constraints, there are clear reasons not to exclude AI. It is anticipated that through currently available AI technology, the impact of employees on financial services can be standardized upwards. It is expected that through any channel, high-quality financial services can be provided to customers. Moreover, AI can comprehensively understand customers and provide customized financial services, and it is expected that such services will be available.

By reducing human errors and increasing the prevention rate of various financial accidents, a safer and more reliable financial system can be established. Now, let me talk about the AI introduction and utilization strategy of Hana Financial Group. The AI strategy that Hana Financial Group considers is first, the customer; second, the customer; third, the customer. The reason for introducing AI is to solve customer problems. Digital AI, as perceived by Hana Financial Group, is a tool for solving customer problems. Recently, rapidly emerging AI technologies are causing unprecedented tension and reshaping the financial market. Hana Financial Group has selected AI as one of the key tasks for leading digital finance, clearly setting the goal of developing and utilizing financial AI for customers and the field based on three directions: enhancing customer value, strengthening AI's impact on the field, and focusing on finance AI for customers and the field. First, reevaluating customer value is crucial, as the approach has been traditionally technology-centered.

We prioritize safe and reliable AI services and customer experience innovation by focusing on problem-solving for AI guests. Next, we have established a control tower for the group to systematically manage the introduction of AI technologies that were previously introduced on a project-by-project basis at the site level, ensuring that site-specific AI technologies are introduced first. Lastly, to swiftly adopt AI enhancement internally, it is easiest to outsource it. However, we, as the Hanwha Financial Group's AI Think Tank, will gradually enhance the group's overall AI technology level by internally driving the rapid adoption of AI. I will now briefly introduce the progress made so far regarding the three AI directions of the Hanwha Financial Group. To realize the mentioned values, in March, we declared the AI Ethics Guidelines. The AI Ethics Guidelines consist of five major points.

It contains principles, and through this, we plan to introduce AI that is safer and more trustworthy. We are also actively forming partnerships between companies to enhance customer value. To focus on the field, we have established an AI control tower within the group and deployed field and technical personnel together like me, concentrating on both technology and its application. We are building governance through the operation of an AI committee attended by executives and practitioners dealing with key issues of related companies, sharing collective intelligence and preventing duplicate investments. Furthermore, we are strengthening the education of field employees, continuously innovating related organizational culture. Lastly, our unique capability is the Hanwha Financial Group's distinctive technology. Hanwha Financial Group operates the Hanwha Financial Technology Institute, a research organization consisting of 57 experts, including doctors and masters. We are conducting joint research on projects required for AI research and practical work with Hanwha technology, focusing on core technologies of financial AI.

We aim to secure a solution that maximizes internal satisfaction, consumer satisfaction, and employee work efficiency to drive the group's digital innovation. This research goal encompasses all aspects related to financial sources, including risk management, asset management, customer experience, and business efficiency, not limited to a specific field. Now, I will introduce AI cases of Hana Financial Group. The cases will be categorized into predictive and generative cases, similar to the ones previously discussed in the UK Finance. The first case is a personalized policy fund recommendation AI model. Various government support funds are available for individual business owners. However, even with the support of branch counter staff, it is challenging to easily recommend policy funds tailored to each customer due to the abundance of policy fund products and their respective support conditions. Therefore, the AI model recommends policy funds that align with the conditions of individual business owner customers.

In the app, we provide customized recommendations to easily access various and complex policy funding loans. Next is the overseas remittance delivery time prediction service. Due to reasons such as time difference and connection with local financial institutions' payment systems, different remittance delivery times occur. Therefore, we predict the delivery time of overseas remittances with AI. After the remittance is completed, we provide real-time progress status until the numerical arrival. Next is a case of introducing generative AI that is currently being promoted. Are you bored now? Let me show you one piece of work. Could you please take a look at the screen? This generative AI was introduced into the customer consultation process at the call center as part of this project. I asked the AI department staff to sketch out the design once. Within a week, they came up with this draft by hand. So, I was deeply moved by the artistic beauty of this piece.

I received it and put it here. In the main part, Hwahwa Financial Group is already operating call and chatbot counseling support. However, there were many limitations in supporting STT refinement, summarizing counseling contents, and systemically converting conversation contents into text. Therefore, AI is introduced to automatically find customer questions during the conversation, show the answers to the counselors first on the screen, so that they can support the conversation with customers while looking at it. After the counseling, the automatically summarized data by AI is linked to show the content to customers when they visit the branch, and if the customer's request has not been resolved, they will be supported through the Encha Market, etc. Lastly, there is the introduction of advertisement copy generation and review-related aspects. As consumer trends change rapidly and digital marketing increases, the absolute quantity and effectiveness of advertisements have become more important. Hwahwa Bank is...

At this moment, 10,000 advertising phrases are being created. If we introduce generative AR, it is said that by learning the content of deliberation and product information, it will automatically generate advertising phrases that comply with the deliberation. It is expected that this will greatly enhance the efficiency of the person in charge. Despite the amazing AI technology that can create videos that feel like reality as in movies, showing practical scenes and windows, there will certainly be improvements with better technology. If you look at the screen now, this picture is called a drum. The drum is a device designed to transport drinking water from a long distance in places where water is scarce and there is no water infrastructure. Of course, there are better technologies such as water pipe connections and groundwater development. However, in countries with poor infrastructure, deriving technologies for water supply via pipelines holds significance. Currently, generative AI is unstable.

Although there are still technical limitations, I believe that it is within the appropriate technology that can be utilized as a tool for solving guest issues. Let me introduce the organization I am in charge of. There are many other financial institutions here. If you look to the left, there is the AI Data Strategy Team of Jigu, responsible for the AI part. Next, there is the Financial AI Department at the bank. If you look to the right, there is a customer care center, which is a call center that you are familiar with. Next is the customer support department at the branch where employees can support customer management, and there is also a team that cares for customers who do not visit the branch. Lastly, there is a data center, a data department, and a MyData business unit within the headquarters. It's a huge organization, with almost a thousand employees. It's quite unique that technology and traditional channels are together. This can explain the AI direction of Hwacha Financial Group.

I will put it directly into the service needed for AI towards guests and employees on the spot. Finally, I will introduce a fun story with this screen. Earlier, I introduced a guest support team in the sales department that manages guests who have only made one transaction in the organization for a year without visiting the store. When these managed guests enter a specific rule, a staff member who manages them appears. So, when you call here, the staff member directly answers the phone and responds to the chat. However, when guests ask questions like, Is this AI? during the chat, why do guests ask these questions a lot? Even guests who only trade digitally communicate, seek trust, and seek human interaction. So, the answer below is like this.

When you answer that it's not, the customer says a lot more and goes a long way. Finance is about people. I think AI will be a tool to help people provide better services. Thank you. [APPLAUSE]

장일호 하나금융지주 금융AI센터 상무 '손님과 현장을 위한 하나금융 AI개발과 활용전략' [2024한국금융미래포럼]

장일호 하나금융지주 금융AI센터 상무 '손님과 현장을 위한 하나금융 AI개발과 활용전략' [2024한국금융미래포럼]

스크립트