MSBA Capstone Final Presentation: Solving Real-World Business Challenges Through Data

The final presentation of the MSBA capstone projects, which focus on solving real-world business challenges through data-driven approaches, was held on Friday, January 16. The event was organized to explore the DTB (Data to Business) model through industry–academia collaboration and to share project outcomes that addressed practical challenges faced by companies using data and analytical techniques. Kyung Sam Park, Center Director, and Jeonghyun Kim, Academic Director of the MSBA program, along with representatives from multiple partner organizations, attended the event both online and offline to review the students’ project results. 

During the presentation, a total of seven teams showcased the outcomes of their capstone projects conducted across various industries and corporate partners. All projects were grounded in a shared guiding question—“How can data be effectively translated into real-world decision-making and execution?”—and focused on designing structures and tools that enable analytical insights to be directly applied in business practice. 

One team, collaborating with LG CNS, proposed a solution to address the “bottleneck between insight and execution” that arises during business process analysis within the rapidly growing BPM (Business Process Management) market. While process intelligence offers significant potential, the team highlighted its limitations in translating insights into execution. To address this challenge, they aimed to develop a system that systematically stores extensive business process assets and integrates AI agents to generate evidence-based Q&A responses and actionable improvement recommendations. In particular, their process modeling based on BPMN (Business Process Model and Notation) and agent design drew considerable attention. 

Another project, conducted in collaboration with LG Household & Health Care, proposed a marketing insight tool leveraging unstructured data. The project focused on enabling faster and more accurate identification of market trends through AI technologies, thereby enhancing decision-making and productivity for marketing practitioners. The team designed an AI agent–based chatbot that allows users to efficiently search for and utilize relevant information within large-scale datasets. 

A team collaborating with PwC proposed an automated corporate performance analysis system utilizing an ontology-based Graph-RAG architecture. Recognizing that time delays in corporate performance analysis directly translate into opportunity costs, the team pointed out the limitations of general-purpose large language models (LLMs) in fully understanding company-specific contexts. To address this challenge, they proposed a multi-agent–based system that structurally pre-learns corporate information, enabling more reliable analysis and faster decision-making. 

Another project, conducted in collaboration with Hyundai Department Store, focused on advancing data-driven ordering decisions in the fresh food category, particularly fruits. Using sales data by date, store, and product from top-performing locations—including the Hyundai Department Store Apgujeong Main Store, Trade Center Store, and Pangyo Store—the team built time-series demand forecasting models and presented a proof-of-concept (PoC) decision support model designed to simultaneously minimize waste rates and stockout rates. A key feature of the project was the application of methodologies tailored to data characteristics, ranging from SQL and statistical techniques to machine learning and deep learning. 

In the Hyundai Motor Blue Members project, the team presented a customized decision-making toolkit designed to strengthen customer retention through promotion- and partner-based strategies. Leveraging integrated Blue Members membership data, the toolkit structures customer status and partner information to identify customers with a high likelihood of behavioral change in the near term, enabling proactive engagement. 

The system also allows for comparisons of partner commission fees and benefit changes from both performance and risk perspectives. By applying different models based on the depth of available customer information, the toolkit recommends optimal partner offerings in advance, supporting more effective and data-driven decision-making. 

Another team, collaborating with Hyundai Motor Securities, presented a project focused on enhancing retirement pension product recommendation services to acquire new IRP (Individual Retirement Pension) customers. In an environment of intensifying competition among banks and securities firms, the team designed a recommendation solution that precisely matches retirement pension products with customers’ investment profiles, with a particular emphasis on downside risk. Targeting second-generation baby boomers as the primary segment, the project placed strong emphasis on reliability and explainability. 

In the project conducted in collaboration with Hyundai Home Shopping, the team proposed a plan to develop recommendation algorithms optimized for Hmall’s business characteristics. By implementing personalized recommendation logic based on customer clustering, the project aimed to increase purchase conversion rates while also designing a structure that considers future expansion into AI-powered shopping agents.

During the feedback session that followed the presentations, partner organization representatives offered a series of positive evaluations. A representative from Hyundai Motor Securities commented, “The portfolio design centered on loss thresholds and the personalized reports demonstrate a level of sophistication rarely seen in existing financial services,” adding that “with appropriate legal review, the solution could be applied directly to real-world operations without difficulty.” 

A representative from Hyundai Home Shopping also noted, “We plan to apply the algorithms developed through this project to Hmall starting in February,” and added that “the project provided practical support in internalizing recommendation systems and expanding AI-based services in the future.” 

The event concluded with closing remarks from Jeonghyun Kim, Academic Director of the MSBA program. He noted that the students’ sustained efforts over the approximately one-year project period, which began last February, were fully reflected in the final outcomes. He also shared that it was particularly impressive to see how students have become increasingly adept at leveraging more advanced technologies with each successive cohort. 

He added, “While the final presentation marks the end of this stage, I encourage you to carry the projects through with a strong sense of responsibility so that the deliverables can be meaningfully applied in real business settings,” and expressed his appreciation to the representatives of the partner organizations for their collaboration. 

The MSBA program is designed around industry–academia collaboration, with a curriculum structured to ensure that data analysis results can be effectively applied in real-world business contexts. The capstone project is a representative example of this approach, in which students define problems based on actual data from partner companies and carry out the full process—from analysis to the design of decision-support tools. The outcomes presented at this final presentation demonstrate that the DTB (Data to Business) education model pursued by the MSBA program is being implemented at a level suitable for practical, real-world application.