Lotte Innovate to Turn On-Device Data Into AX Through Ontology, Builds Lotte Department Store's 'Brand AI'
TECHWORLD ·
✦ AI Summary
Lotte Innovate announced the launch of Lotte Department Store's brand data analysis system, "Brand AI."
Brand AI targets about 4,000 brands, connecting and structuring retail-site data such as sales, customer and purchasing information, and analyzing brands from a relational perspective.
The analysis results are visualized as 2D and 3D network graphs and radial charts, supporting MD decisions related to store openings and closures and brand sourcing.
Lotte Innovate announced the launch of Lotte Department Store's brand data analysis system, "Brand AI." Brand AI is an AI work platform powered by AI and data technologies and targets about 4,000 brands.
The system connects and structures retail-site data such as sales, customer and purchasing information, and applies an ontology-based data framework that reflects business context. Lotte Innovate built a method that links brand-specific sales, customer characteristics and related purchasing data in a graph structure.
The core of Brand AI is analyzing brands from a relational perspective rather than as individual data points. In addition to individual figures, Brand AI analyzes relationships and characteristics among brands, and uses that foundation to examine brand similarity and related-purchase relationships.
The analysis results are visualized as 2D and 3D network graphs. This is designed to help managers intuitively understand relationships among brands and their positions in the market, supporting MD decisions related to store openings and closures and brand sourcing.
The company also implemented a "Brand DNA" feature for comparing brand characteristics. Brand DNA is a structured format of 6 key indicators, including brand growth potential, customer characteristics and sales scale, with each indicator normalized on a 0 to 100 scale. This makes it possible to compare brand characteristics.
The analysis results are provided in a radial chart. In addition, for indicators with large brand-by-brand variance, such as sales scale and average spending per customer, the company applied data transformation and outlier adjustment. The company said these steps were intended to improve comparability.
The company also applied large language model (LLM) and retrieval-augmented generation (RAG) technologies to its generative AI functions. AI searches and references relevant data, and automatically generates key brand characteristics and insights within a predefined standard keyword framework.
Generated insights can be managed in a structured format. In addition, sensitive data such as sales revenue and number of customers was used by converting actual figures into grades and ratios instead of raw numbers. The company said it took data security into account by using grade and ratio conversions.
A service foundation has been established to ensure stable use of Brand AI in actual business operations. To that end, integration with internal systems, data management, and the operating environment were implemented together.
Seung-Nyeong Jeon, executive vice president of Lotte Innovate's D&AX division, said there was close collaboration between Lotte Department Store and Lotte Innovate from the early planning stage of Brand AI, and that combining Lotte Department Store's MD expertise and field experience with Lotte Innovate's ontology and AI technologies was meaningful in building a service with high practical utility. He added that the company plans to continue expanding industry-specific data and AI services and to support client companies in innovating their work and strengthening competitiveness.
Source: TECHWORLD · Lee Kwang-jae
Original: https://www.epnc.co.kr/news/articleView.html?idxno=406340
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.