ATradGPT: leveraging agentic ai to enhance stock trading education and accessibility in capital markets

Abstract

Building a good investment portfolio within the Colombo Stock Exchange (CSE) is a complex process with a steep learning curve, often requiring access to real-time information, understanding of market dynamics, and adherence to trading regulations. However, this information and material are not available at a single location, causing inconvenience to users. Furthermore, even if it is accessible, it is often a very demanding task for a single person to process the said material and make informed decisions. To democratize investing by facilitating access to educational resources and streamlining the trading process, a Conversational Intelligent Investment Advisor has been developed. Leveraging open-source models with the RAG architecture, Agentic workflows and Langchain framework, the intelligent advisor integrates with the ATrad application and guides users through the complete process of investing.