Augmenting Fundamental Analysis with Large Language Models: A RAG-Based System for Generating Investor Briefs

Short summary

This study explores using LLMs with a RAG pipeline to augment fundamental analysis by processing company reports, macroeconomic data, and SEC EDGAR filings through the GPT-4o API. The system generated automated investor briefs for 9 companies over 4 weeks, incorporating domain knowledge based on Kitchin cycles. Nine individual investors evaluated the usefulness of the LLM-generated briefs for data analysis, providing real-world feedback on the approach.

• •RAG-based system using GPT-4o to generate investor briefs from SEC filings and macroeconomic data
• •Processed reports for 9 companies over 4 weeks with Kitchin cycle domain knowledge
• •Nine individual investors evaluated the automated briefs for practical usefulness

Generated with AI, which can make mistakes.