K Bank Publishes Research Paper on 'AI-Based Personalized Recommendations' in Domestic Academic Journal

Paper on AI-Based Personalized Recommendation System Published in JKDAS Journal
Continuing to Accelerate Financial Innovation through AI Technology

K Bank has demonstrated the innovation and excellence of its AI-based personalized recommendation system by presenting the theoretical basis for its strategic design and performance.

On July 14, K Bank announced that its paper, "Strategic Design of AI-based Recommendation Systems and Analysis of Changes in User Experience: Financial App Experimentation through MLOps Automation," which covers the 'personalized' K Bank app reflecting customer behavior patterns, has been published in the domestic data analysis journal JKDAS (Journal of the Korean Data Analysis Society).

JKDAS is a journal published by the Korean Data Analysis Society (KDAS) and is one of the major domestic academic journals listed in the Korea Citation Index (KCI), actively covering statistical data analysis theory and applied research.

This study empirically analyzed the impact of AI technology on customer behavior change, user experience, and corporate profitability, focusing on the personalized recommendation system applied to the K Bank app. It is considered a meaningful attempt as a study on AI model personalization strategies within the financial sector, rather than in industries such as commerce or OTT where personal recommendations are more common.

In particular, to design a model optimized for the financial industry, K Bank conducted Focus Group Interviews (FGI) with representatives from various financial sectors within the bank, including lending and deposits.

Through this process, K Bank was able to precisely identify customer types and behavior patterns and apply these results from the initial stages of AI model development. This approach goes beyond simple technology-driven recommendations by reflecting the characteristics of financial consumers and enhancing both predictive performance and operational stability.

This recommendation system is implemented based on MLOps (Machine Learning Operations), enabling real-time detection and analysis of behavioral data such as customer preferences and time spent within the app. A key feature is the creation of an automated process that allows the system to continuously learn from the analysis results and reapply them to the system.

A K Bank representative stated, "We plan to ultimately evolve the self-developed system based on AI technology into an AI Agent framework to provide customers with even more sophisticated financial services," adding, "We will continue to focus on leading AI-based financial services and transform into an 'AI Powered Bank.'"

Meanwhile, since the beginning of this year, K Bank has accelerated AI financial innovation by introducing a private LLM (Large Language Model), building AI automation systems to improve internal work efficiency, and laying the groundwork for expanding customer-facing AI services. In particular, the company has invested in AI and cloud infrastructure at a level about three times higher than last year, including expanding GPU servers.