CrowdWorks-Yuanta Develop Context-Aware AI Agent for Stock Analysis

Crowdworks and Yuanta Securities have collaborated to introduce 'Yuanta AI Agent,' an innovative AI agent service for stock analysis. Designated as an innovative financial service by the Financial Services Commission last January, this service goes beyond simple information retrieval to deeply understand the user's question context and provide comprehensive answers. 'Yuanta AI Agent' provides a vast amount of information that investors need in real-time, including corporate financial information, market outlooks, stock prices, industry trends, and legal regulations. Unlike existing LLM (Large Language Model)-based AI chatbots, it precisely analyzes the user's question intent to search appropriate databases, and is characterized by generating richer, real-time, and highly reliable answers. The core of this service lies in the Agentic RAG technology of Crowdworks' agentic AI development solution, 'Alpy'. Alpy RAG provides an advanced framework that enables the agent to independently formulate optimal search strategies and find the most accurate answers through multiple steps. Based on this technology, 'Yuanta AI Agent' selects the most relevant information from various databases, such as financial data and market trends, and synthesizes it to provide investors with highly reliable analysis results. In practical application examples, when querying a specific company's ROE (Return on Equity), it not only provides the data but also explains the concept of ROE in detail, and proactively offers additional information such as related financial indicators by anticipating subsequent user questions. Furthermore, by effectively utilizing charts, graphs, and images, it supports comparing financial statements between companies and classifying stock-related news as positive or negative, thereby enhancing information comprehension. Service management functions have also been significantly enhanced. Through the administrator page, the AI agent environment can be efficiently managed by real-time monitoring of user queries, token usage, and costs associated with LLM usage. In addition, it detects sensitive information or security violations that may be included in user queries, and provides various management tools, such as setting guidelines for delivering compliance-aligned answers. Kim Woo-seung, CEO of Crowdworks, stated that "through Alpy technology, we have overcome the limitations of existing AI chatbots and implemented a more sophisticated and reliable stock analysis AI service." He added that Crowdworks will establish itself as a key partner leading AI innovation in the financial sector through future collaborations with various financial institutions.