Ziyuan Wang¹ and Carlos Gonzalez², ¹USA, ²California State Polytechnic University, USA

Historically, comprehensive stock market insights were largely inaccessible, requiring significant effort from individuals. This paper presents StockBot AI, a mobile application designed to democratize financial engagement by providing personalized, real-time investment insights through a conversational AI [1]. The system integrates live stock and news data from sources like Alpha Vantage, synthesizing it via a proprietary Bullish-Bearish scoring model and feeding this contextually rich information to a Large Language Model. Key challenges included ensuring data validity and accuracy from diverse sources, preventing LLM hallucinations, and maintaining the scoring system's relevance in stochastic markets [2]. These were addressed by redundant data pipelines, advanced RAG techniques, and continuous model calibration. Through automated comparative experimentation, StockBot AI demonstrated superior performance in delivering contextually relevant insights, particularly due to its unique integration of news sentiment. Ultimately, StockBot AI offers a user-friendly, one-place solution, empowering individuals to confidently navigate and responsibly participate in the economy.

Stock Market, Data Analytics, Artificial Intelligence, Software