Ontology-Based Field Data AX Drives Lotte Department Store's 'Brand AI' Buildout

[Financial News] LOTTE INNOVATE said on the 1st that it has built 'Brand AI,' a brand data analysis system for Lotte Department Store, using artificial intelligence and data technologies. Earlier, Lotte Department Store's AI Studio designed a field-oriented AI agent architecture and directly implemented a mock-up that practitioners could experience through vibe coding, clearly defining the technical requirements.
Brand AI is an AI work platform that applies an ontology-based data framework to connect and structure a wide range of retail data, including about 4,000 brands, as well as sales, customer and purchasing information, in line with business context. It supports merchandise planners' decisions on store openings and closures, brand sourcing and other tasks by analyzing not only individual figures but also relationships and characteristics among brands.
The core of 'Brand AI' is analyzing brands from the perspective of 'relationships' rather than as isolated data points. LOTTE INNOVATE linked each brand's sales, customer characteristics and related purchase data in a graph structure, then used it to analyze brand similarity and related-purchase relationships. The results are visualized as 2D and 3D network graphs, allowing staff to intuitively understand relationships between brands and their positions in the market.
It also developed a 'Brand DNA' feature for comparing brand characteristics. Brand DNA structures six key indicators, including growth potential, customer characteristics and sales scale, and normalizes each indicator on a 0 to 100 scale so that brand traits can be compared. The results are presented in a radar chart, and indicators with large brand-to-brand gaps, such as sales scale and average spending per customer, were adjusted through data transformation and outlier correction to improve comparability.
For its generative AI functions, the company applied large language model (LLM) and Retrieval-Augmented Generation (RAG) technologies. The AI was configured to search and reference relevant data and automatically generate key brand characteristics and insights within a predefined standard keyword framework. The generated insights were managed in a structured format, while sensitive data such as sales and customer counts were converted into grades or ratios instead of actual figures to address data security concerns.
In addition, the company built integration with internal systems, data management and an operating environment, laying the foundation for Brand AI to be used stably in actual business operations. LOTTE INNOVATE carried out the entire process, from data construction and AI analysis to service implementation and operations.
Jeon Soong-nyeong, Managing Director of LOTTE INNOVATE's D&AX business division, said, "Brand AI shone through the close collaboration between Lotte Department Store and LOTTE INNOVATE from the early planning stage," adding, "This project is highly meaningful because it combined Lotte Department Store's MD expertise and field experience with LOTTE INNOVATE's ontology and AI technologies to build a service with stronger practical usability."
[email protected] Yun Ji-an Reporter