FinSight AI: A Multi-Agent Retrieval-Augmented Generation Framework for Real-Time Financial Analysis and Investment Assistance

FinSight AI offers an artificial intelligence framework designed for financial market analysis and investment decision support. The platform brings together real-time market data, technical analysis tools, financial news and social media sentiment. It also uses Retrieval-Augmented Generation (RAG) and Large Language Models (LLM) to enable natural-language conversation. All these pieces work together in one system to help users get financial insights through simple human-like interactions.   The review of existing research shows that combining market data with qualitative inputs—like news articles and social media trends—can give a fuller picture of market movements. Recent studies also point to potential in using Large Language Models, RAG and multi-agent systems for financial tasks. FinSight AI builds on this by combining financial analysis tools with LLM-based reasoning. This allows the system to process kinds of financial data and present results in a way that is easy to understand.   Beyond core analysis FinSight AI adds useful features. These include support for Systematic Investment Plans (SIP) tracking of Initial Public Offerings (IPO) and a paper trading environment. These extras help users explore investment ideas in a simulated setting. The use of RAG ensures that the system can bring in external information when answering questions. At the time technical indicators and sentiment signals add both quantitative and qualitative layers to the analysis.   Overall FinSight AI shows how modern AI technologies can be combined with data to build a smart and user-friendly decision-support platform.. The system’s reliability depends on the quality,availability and up-, to-date nature of external financial data. It also depends on the limits of what LLMs can generate. Therefore FinSight AI should be seen as an educational tool—not as a promise of investment success. Future improvements can focus on forecasting methods, more accurate retrieval and sentiment analysis, clearer explanations, wider financial datasets and stronger validation processes.

[1]   M. Luckianto and A. A. S. Gunawan, “MARAG-Fin: An Intelligent Multi-agent RAG-LLM Architecture Integrating Financial News Sentiment and Time Series Data for Data-driven Trading Decision-making,” International Journal of Intelligent Engineering and Systems, vol. 19, no. 1, pp. 738–748, 2026.   [2]  S. Das, S. Ghosh, A. Tiwari, R. E. Z. M. Lynghoi, S. Saha, Z. Murad, and A. Maurya, “Presenting an Order-Aware Multimodal Fusion Framework for Financial Advisory Summarization With an Exclusive Video Dataset,” IEEE Access, vol. 13, pp. 48364–48381, 2025.   [3]   K. Liagkouras and K. Metaxiotis, “A Hybrid Long Short-Term Memory with a Sentiment Analysis System for Stock Market Forecasting,” Electronics, vol. 14, no. 2753, pp. 1–18, 2025.   [4]    A. Jadhav and V. Mirza, “Large Language Models in Equity Markets: Applications, Techniques, and Insights,” Frontiers in Artificial Intelligence, vol. 8, no. 1608365, pp. 1–15, Aug. 2025.   [5]   N. Srivatsan, “Using Large Language Models to Predict Stock Price: A Hybrid Model Combining Social Media Sentiment and Market Data,” National High School Journal of Science, pp. 1–12, Aug. 2025.