Title:

Building an AI Investment Advisor: A Multi-Agent System for Real-Time Stock Analysis and Portfolio Recommendations

Investing is a powerful tool for building financial security. When starting out, retail investors struggle with interpreting market data, managing risk, and making timely decisions. Existing tools can be helpful but have barriers like manual analysis and recommendations which are not personalized to the individual users. This project introduces an AI powered investment advisor that extracts real-time stock data while adapting to the user's goals and current investments. Using stock data, news sentiment, and user input, this investment advisor will generate recommendations, supporting financial decision-making towards their financial goals. This investment advisor utilizes a full-stack architecture which gathers the user’s investment goal and current portfolio (ticker, shares, average buy price, and risk tolerance). The investment advisor then takes this data and searches the Alpha Vantage API for the current stock prices and NewsAPI for the sentiment on the stock which then creates a confidence score on each stock using a rule-based agent. This info is gathered then given to Groq which accesses the Llama LLM which is accessed through LangChain to generate an explanation for the recommendations. Results demonstrate that the investment advisor is able to provide consistent and actionable recommendations to the user with confidence scores ranging from 0-1. Portfolios along with updates from the generated recommendations are saved within a JSON-based memory.

Name

Comment*

Comments are viewable only by submitter

Captcha*