标题： 结构化非结构化：一种用于提取和查询财务KPI和指导的多智能体系统 显示英文标题

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

摘要： 从非结构化的财务文件中提取结构化和定量的见解在投资研究中是至关重要的，但仍然耗时且资源密集。 实际中的传统方法严重依赖于人力密集的手动流程，限制了可扩展性并延迟了研究流程。 在本文中，我们提出了一种高效且可扩展的方法，用于准确地从非结构化的财务文档中提取定量见解，利用由大型语言模型组成的多智能体系统。 我们提出的多智能体系统包括两个专业代理： \emph{提取代理}和\emph{文本到SQL代理}。 \textit{提取代理}自动从非结构化的财务文本中识别关键绩效指标，标准化其格式并验证其准确性。 另一方面，\textit{文本到SQL代理}从自然语言查询生成可执行的 SQL 语句，使用户能够准确访问结构化数据，而无需熟悉数据库模式。 通过实验，我们证明了我们的系统能够准确地将非结构化文本转换为结构化数据，并实现关键信息的精确检索。 首先，我们证明了我们的系统在将财务文件转换为结构化数据方面达到了约 95% 的准确率，达到了通常由人工标注者获得的性能水平。 其次，在对检索任务的人工评估中——其中使用自然语言查询从结构化数据中搜索信息——91% 的回答被人类评估者评为正确。 在两项评估中，我们的系统在各种类型的财务文档中都能很好地泛化，始终提供可靠的表现。 显示英文摘要

摘要： 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.