KakaoBank Enhances Financial Accessibility with Alternative Credit Assessment and AI

Supplying 1 Trillion Won in Loans to Mid-Credit Customers
Through Alternative Credit Assessment Using Non-Financial Data
Employees Freely Utilize Generative AI
To Launch Customer-Facing Services

KakaoBank, an internet-only bank, announced that it is enhancing financial accessibility through its unique strategies. The bank is expanding loans to individuals with mid- to low-credit scores by leveraging alternative credit assessment models based on non-financial data, and is creating a safer financial environment through generative artificial intelligence (AI) services.

KakaoBank Expands Inclusive Finance Using Alternative Data

On the afternoon of the 29th, KakaoBank held the "2025 KakaoBank Connect" event and stated that it has been developing alternative credit assessment models to more accurately evaluate the creditworthiness of mid- to low-credit customers and those with limited financial histories (thin filers). KakaoBank first began utilizing alternative data in 2019, when it used telecommunications data to give additional credit points to loan applicants. In 2021, the bank incorporated non-financial information into its proprietary credit assessment model developed using machine learning (ML), and began actively using it to process loans for mid- to low-credit customers.

The alternative credit assessment model was first introduced in 2022. By utilizing 18 million pseudonymized and combined data points from companies such as Lotte Members and Kyobo Bookstore, KakaoBank developed "KakaoBank Score," the industry’s first independent alternative credit assessment model. In 2023, the bank established a "Small Business Specialized Credit Assessment Model" for sole proprietors, lowering the loan threshold for sectors with limited financial access, such as the restaurant industry. Currently, KakaoBank operates five alternative credit assessment models: "KakaoBank Score" and "KakaoBank Trust Score" for individuals, and specialized scores for restaurants, service industries, and online sellers for sole proprietors.

Jin-Hyun Cho, Head of Credit Risk Modeling Team at KakaoBank, is explaining KakaoBank's alternative credit scoring model. KakaoBank

In particular, the KakaoBank Score is used to grant additional loan approvals to high-quality mid-credit customers who were initially rejected by conventional models. For example, if a customer who was denied a loan by the basic credit assessment model frequently checks their KakaoBank group account, receives many gifts through "Kakao Gift," often uses taxis on weekends, or purchases many foreign language books from Kyobo Bookstore, they may be approved for an additional loan. The total amount of loans additionally executed for mid-credit products through this method reached approximately 989.3 billion won, accounting for 13% of all mid-credit product loans. The amount of additional loans executed for thin filers using the KakaoBank Score was about 269.3 billion won, representing 27% of all loans processed with the KakaoBank Score.

The KakaoBank Score, created from alternative data, has also contributed to expanding inclusive finance. Comparing the top 30% of standard credit scores, typically used by banks and developed by credit rating agencies, with the top 30% of KakaoBank Scores, only 1.8% of the former are thin filers, while 9.7% of the latter are thin filers. The same applies to low-income individuals: they make up only 5.3% of the top 30% by standard credit score, but 10.6% by KakaoBank Score. Jin-Hyun Cho, Head of Credit Risk Modeling Team at KakaoBank, stated, "We will continue to broaden the foundation of our alternative credit assessment models, not just for KakaoBank but for the entire nation, to realize truly inclusive finance that allows everyone to access fair financial services."

Creating a Safe Financial Environment with Generative AI

KakaoBank is also working to create a safer financial environment by implementing generative AI services throughout its application. Four key services include AI-based smishing message detection. The AI smishing message detection service analyzes the content of messages to determine if they are smishing attempts and provides the rationale for its judgment. The AI search feature delivers highly accurate answers based on the latest data when customers ask questions such as "What is the current savings interest rate?" The AI financial calculator understands customer queries and automatically applies the necessary calculations. The in-app chatbot independently comprehends the intent behind customer inquiries and responds naturally, with a failure rate of less than 1%.

Jaewook Lee, Head of AI Customer Service Development Team at KakaoBank, is explaining KakaoBank's generative AI service. KakaoBank

Such services are made possible by the "AI Playground," where employees can naturally use and experiment with generative AI, and the Financial Technology Research Institute, which conducts ongoing research. The AI Playground is an internal platform that allows staff to directly experiment with and develop AI, enabling anyone to understand and use AI in their daily work. For example, tools like "Recommend a restaurant in Pangyo for a company dinner" are work support bots created by employees using generative AI and shared within the company for others to use freely. The Financial Technology Research Institute was designated as an innovative financial service in 2020, allowing for a relaxed research and development environment with eased network separation regulations. Through this, the institute has conducted research on new technologies such as AI, authentication, and big data, filed 169 patents (111 domestic and 58 overseas), and presented more than 16 papers at domestic and international conferences.

Jaewook Lee, Head of AI Customer Service Development Team at KakaoBank, stated, "AI may not be able to solve every problem, but it is the most useful tool to help customers use financial services more easily and safely. We will continue to solve financial consumer issues through technology and create an environment where everyone can enjoy using financial services."

This content was produced with the assistance of AI translation services.