This dataset contains 140 measured responses across five K-beauty brands, seven AI engines, and four prompts. We publish the results and measurement conditions so others can reproduce and compare Korean-language AI search visibility.
The five brands averaged 93.2/100 in Share of Voice (SoV). Brand mentions were observed in responses from four global engines (ChatGPT, Claude, Perplexity, and Gemini) and two Korean engines (HyperCLOVA X and Naver). This sample does not establish why they appeared.
Daum mentioned the five brands in 55% of measured answers on average, below the other engines measured. Anua reached 75%, showing variation by brand. The dataset alone cannot establish the cause.
Anua scored 96/100 in this sample. Identifying whether distribution, reviews, or web documents influenced the result would require source analysis and repeated measurements.
Use the dataset to evaluate Korean-language LLMs, examine GEO methods, or analyze K-beauty. Field definitions and measurement conditions are included in the downloadable files.
Planned expansion: 50 K-beauty brands in v0.5 (September 2026), then 200 brands across K-beauty, K-fashion, and K-content in v1.0 (December 2026). A joint R&D publication with Naver is under consideration; it is not confirmed.
Start in 3 minutes
Just your email, done in 3 minutes. No card required.