MegazoneCloud said on the 2nd that it built an AI shopping agent that recommends personalized golf equipment for Golfzone Commerce's online mall Golping.
The service uses more than 500,000 fitting data points accumulated at Golfzon screen golf facilities to propose clubs tailored to each user's swing characteristics. The data used for recommendations consists of measured records that meet conditions affecting accuracy, such as the continuity of swing information.
The core is a multi-agent system. An AI fitting agent group analyzes the user's swing mechanism to create a golfer profile, and a product recommendation agent group infers suitable equipment based on it. The results of the two groups are combined to present a single recommendation to the user.
MegazoneCloud implemented this system on Amazon Bedrock Agents Core. Agents Core coordinates the workflows of multiple agents. The model applies Claude Sonnet 4.5 on Amazon Bedrock, and caching technology is used for repeated queries to improve response speed and expense efficiency. To ease concerns over handling sensitive data, it was designed to operate in the Seoul Region.
Golfzone Commerce has been offering an AI fitting service on Golping since May 7. Users can check three club recommendations matched to their swing traits, along with the basis for the recommendations and the expected improvement effects. Golf balls are also proposed by reflecting play style, preferred feel at impact, and spin requirements. Members without swing data can receive similar recommendations by answering a questionnaire.
The AI shopping agent also provides real-time inventory for recommended products and information on nearby Golfzon Market stores based on the user's location. It links online recommendations to actual purchases and visits to offline stores.
Golf gear is a high-involvement product in which satisfaction varies widely—even for the same item—depending on a user's body type, swing path, shot shape, and preferred feel at impact. The industry says AI fitting can reduce the uncertainty of online purchases and serve as a catalyst to shift equipment recommendations—once centered on consultations—into a data-driven service.
Vice President Yoo Hyeong-rim of MegazoneCloud said, "This is a case showing that when corporate data combines with agentic AI, it can lead to business results."