Advancing ESG Intelligence: An Expert-level Agent and Comprehensive Benchmark for Sustainable Finance

Unofficial AI-generated summary based on the public title and abstract. Not an official translation.

🔬研究者:Provides a benchmark and agent architecture for AI-driven ESG analysis, advancing research in vertical LLM applications.

🏢実務担当者:Corporates and auditors can leverage ESGAgent to automate ESG report analysis and generate verifiable insights.

🏛政策担当者:Offers insights into how AI can enhance ESG data reliability, potentially informing future regulation on AI-assisted auditing.

Environmental, social, and governance (ESG) criteria are essential for evaluating corporate sustainability and ethical performance. However, professional ESG analysis is hindered by data fragmentation across unstructured sources, and existing large language models (LLMs) often struggle with the complex, multi-step workflows required for rigorous auditing. To address these limitations, we introduce ESGAgent, a hierarchical multi-agent system empowered by a specialized toolset, including retrieval augmentation, web search and domain-specific functions, to generate in-depth ESG analysis. Complementing this agentic system, we present a comprehensive three-level benchmark derived from 310 corporate sustainability reports, designed to evaluate capabilities ranging from atomic common-sense questions to the generation of integrated, in-depth analysis. Empirical evaluations demonstrate that ESGAgent outperforms state-of-the-art closed-source LLMs with an average accuracy of 84.15% on atomic question-answering tasks, and excels in professional report generation by integrating rich charts and verifiable references. These findings confirm the diagnostic value of our benchmark, establishing it as a vital testbed for assessing general and advanced agentic capabilities in high-stakes vertical domains.

• semanticscholar https://doi.org/10.48550/arxiv.2601.08676first seen 2026-05-06 00:29:40