Amorepacific Establishes AI-Driven Research Platform Leveraging 70 Years of R&D Assets

Amorepacific has established a research-focused generative AI platform based on over 70 years of accumulated research and development assets. The company, in collaboration with KT, has initiated the 'Data Highway' project to redesign research assets so that they can be comprehended and utilized by AI.

This project has led to the development of an AI-centric service known as 'AI Assistant LEMON (Lab Efficiency Mode ON).' Since founding the first cosmetics research lab in South Korea in 1954, Amorepacific has gathered extensive research and development assets, including raw material data, formulation data, research reports, patents, and academic papers. However, these data have been dispersed across various systems, presenting challenges in locating and utilizing the necessary information, and limiting the effective use of AI.

To address these challenges, Amorepacific, supported by its internal AX organization, has worked with KT on the 'Data Highway' project since last year. The two entities have jointly designed and developed a platform and AI services optimized for AI use, creating an 'AI Data Hub' that integrates and standardizes millions of structured and unstructured data points, including raw material data, formulation data, and research reports.

This foundation has enabled the development of a large language model (LLM) service specifically for research: 'AI Assistant LEMON.' Researchers can now utilize LEMON to quickly and efficiently search, analyze, and apply extensive research data across various tasks, including research and development, production, quality control, and regulatory compliance. The AI system also allows for diverse exploration and discovery during the research process.

Results from the beta test of 'AI Assistant LEMON' conducted with Amorepacific's R&I Center researchers indicated a significant reduction in previously lengthy tasks. The time taken for product planning and regulatory review—which used to take approximately 12 days—has been shortened to about 5 minutes. Additionally, the time required to review the research history and product applicability of specific raw materials has decreased from around 6 days to approximately 5 minutes.

Through this project, Amorepacific aims not only to implement AI services but also to innovate the entire research environment. Plans are in place to expand digital transformation tasks, including introducing 'AI Agents,' thereby reducing repetitive information-seeking tasks for researchers and enabling them to concentrate on creative research activities, such as formulating new hypotheses and technical innovations.

Seo Byung-hwi, CTO of Amorepacific's R&I Center, stated, “Good AI becomes a research partner only when it meets good data. This project is significant as it shows Amorepacific is leading the transition to a research environment appropriate for the AI era. In the future, the AI platform will handle repetitive exploration and analysis, allowing researchers to focus on creative questions and innovations. The Data Highway and LEMON will serve as a starting point for creating a new research environment.”