Musinsa said on Wednesday it has launched an "AI review summary" service that uses artificial intelligence (AI) technology to make it easier to understand tens of thousands of product reviews.

The AI review summary service analyzes and refines about 21 million reviews from actual users accumulated over the past two years on Musinsa Store. The company said it expects the feature to provide customers with a more efficient fashion shopping experience.

Musinsa holds about 11.52 million reviews a year. It generates an average of about 960,000 purchase records each month. Musinsa placed the AI review summary function prominently at the top of the review section on product detail pages (PDP) so customers can quickly filter key information they need, such as fit and material, from large amounts of data.

The service reflects the characteristics of the fashion category in detail. With the "pros and cons summary" function, it plans to disclose negative factors such as points to note along with positive evaluations to ensure objectivity of information. It also provides keyword summaries optimized for each clothing category. For example, for pants it focuses on size, fit and stretch, while for outerwear it emphasizes design, warmth and material.

Musinsa built a feedback loop that allows customers to rate AI summaries directly as "helpful" or "not helpful". Detailed user opinions collected in the rating process will be used as base data to refine the AI model's sentence generation ability and improve the accuracy of answers in the future.

A Musinsa official said that about 80% of users clicked "helpful" during the first 5 days after the service launch, showing the usefulness of the AI summary information. The official said the company will reflect user feedback in real time because it is currently in the beta test stage, continuously upgrade the model and expand AI functions to improve shopping convenience.