머큐리프로젝트 Co., Ltd. announced that it has integrated its AI agent infrastructure “MERCURY X” with ChatGPT and Claude via the Model Context Protocol (MCP), enabling users to call and use it directly within those environments.

With this integration, users no longer need to navigate complex menus or generate reports. By asking a simple one-line question in the chat window, such as “Are there any products at risk of stockout within 30 days?”, MERCURY X narrows down and displays at-risk products based on available indicators such as recent purchase concentration, turnover rate, and rank changes, and also presents the grounds for why they are considered risky.

The key point of this MCP integration is the division of roles. ChatGPT and Claude function as conversational interfaces (inputs) that receive users’ questions in natural language, while the MERCURY X agent system (the “brain”) is responsible for actually calculating the data and making judgments.

Users simply ask questions as if speaking to a person. For example, when they ask, “Which products should be pushed more?”, “Where should this product be sold to generate more profit?”, or “How much inventory is safe to hold?”, MERCURY X is designed to connect the entire flow from analysis → strategy → execution.

A MERCURY Project official stated, “While most SaaS solutions use LLMs only as conversational interfaces, MERCURY X is designed so that the agent system is responsible for judgment and execution while minimizing the use of LLMs,” adding, “This MCP integration shows that AI is evolving from a ‘supporting tool’ into a structure that actually performs work.”

Jeong Jin-moon, CEO of MERCURY Project, stressed that many brands are facing declining profits despite sales growth, due to rising advertising costs and fees, inventory burdens, and discount competition. Ultimately, the key is not “selling more” but “making profit-generating decisions at the right time.”

MERCURY X provides analysis and decision-making information for overall brand operations, including SKU optimization, detection of stockout and inventory risks, marketing budget allocation, and product, pricing, and channel operations. With this MCP integration, e-commerce operators can reduce the time spent searching for metrics and creating reports, and instead implement a Q&A-style UX where they confirm decision rationales and priorities instantly through questions in the chat window, allowing them to focus on decision-making and execution.

The company stated, “Since its launch in September last year, MERCURY X has been used by around 25 medium and large cosmetics, fashion, and F&B brands, and is being utilized for operational decisions that directly affect profits, such as pricing, inventory, product, and advertising operations,” adding, “We are advancing MERCURY X as an AI agent infrastructure and are currently in discussions with domestic and overseas investors.”

Popular News