Structuring the Unstructured: A Multi-Agent System for Extracting and Querying Financial KPIs and Guidance

TL;DR

This paper introduces a multi-agent system utilizing large language models to efficiently extract and query financial KPIs from unstructured documents, achieving high accuracy and scalability in financial data analysis.

Contribution

The paper presents a novel multi-agent system with specialized agents for extracting KPIs and converting natural language queries into SQL, significantly improving scalability and accuracy over manual methods.

Findings

Achieves approximately 95% accuracy in data extraction from financial filings.

91% of retrieval responses rated correct by human evaluators.

System generalizes well across different financial document types.

Abstract

Extracting structured and quantitative insights from unstructured financial filings is essential in investment research, yet remains time-consuming and resource-intensive. Conventional approaches in practice rely heavily on labor-intensive manual processes, limiting scalability and delaying the research workflow. In this paper, we propose an efficient and scalable method for accurately extracting quantitative insights from unstructured financial documents, leveraging a multi-agent system composed of large language models. Our proposed multi-agent system consists of two specialized agents: the \emph{Extraction Agent} and the \emph{Text-to-SQL Agent}. The \textit{Extraction Agent} automatically identifies key performance indicators from unstructured financial text, standardizes their formats, and verifies their accuracy. On the other hand, the \textit{Text-to-SQL Agent} generates executable SQL statements from natural language queries, allowing users to access structured data accurately without requiring familiarity with the database schema. Through experiments, we demonstrate that our proposed system effectively transforms unstructured text into structured data accurately and enables precise retrieval of key information. First, we demonstrate that our system achieves approximately 95\% accuracy in transforming financial filings into structured data, matching the performance level typically attained by human annotators. Second, in a human evaluation of the retrieval task -- where natural language queries are used to search information from structured data -- 91\% of the responses were rated as correct by human evaluators. In both evaluations, our system generalizes well across financial document types, consistently delivering reliable performance.