@inproceedings{zhao-etal-2026-advancing,
title = "Advancing {ESG} Intelligence: An Expert-level Agent and Comprehensive Benchmark for Sustainable Finance",
author = "Zhao, Yilei and
Zhang, Wentao and
Xiao, Lei and
Zheng, Yandan and
Liu, Mengpu and
Lim, Wei Yang Bryan",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.621/",
pages = "14256--14290",
ISBN = "979-8-89176-459-0",
abstract = "Environmental, social, and governance (ESG) criteria are essential for evaluating corporate sustainability, yet professional ESG analysis is hindered by data fragmentation and complex multi-step workflows. Existing ESG benchmarks rely on LLM-generated questions that often suffer from semantic ambiguity, limiting their diagnostic reliability. We introduce ESGAgent, a hierarchical multi-agent system with a domain-oriented toolset for automating end-to-end ESG workflows, and a three-level benchmark derived from 310 corporate sustainability reports, featuring expert-sourced questions with verified, unambiguous answers at Levels 1-2 and full report generation at Level 3. Evaluations across 20 state-of-the-art closed-source LLMs show that ESGAgent achieves 84.15{\%} accuracy on Levels 1-2, surpassing the strongest baseline by 3.26 points, and produces Level 3 reports with higher factual consistency and analytical depth than commercial deep research systems, validated by both multi-LLM judging and human expert evaluation, showing a robust alignment (Pearson $r=0.551$)."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhao-etal-2026-advancing">
<titleInfo>
<title>Advancing ESG Intelligence: An Expert-level Agent and Comprehensive Benchmark for Sustainable Finance</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yilei</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wentao</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lei</namePart>
<namePart type="family">Xiao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yandan</namePart>
<namePart type="family">Zheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mengpu</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wei</namePart>
<namePart type="given">Yang</namePart>
<namePart type="given">Bryan</namePart>
<namePart type="family">Lim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Environmental, social, and governance (ESG) criteria are essential for evaluating corporate sustainability, yet professional ESG analysis is hindered by data fragmentation and complex multi-step workflows. Existing ESG benchmarks rely on LLM-generated questions that often suffer from semantic ambiguity, limiting their diagnostic reliability. We introduce ESGAgent, a hierarchical multi-agent system with a domain-oriented toolset for automating end-to-end ESG workflows, and a three-level benchmark derived from 310 corporate sustainability reports, featuring expert-sourced questions with verified, unambiguous answers at Levels 1-2 and full report generation at Level 3. Evaluations across 20 state-of-the-art closed-source LLMs show that ESGAgent achieves 84.15% accuracy on Levels 1-2, surpassing the strongest baseline by 3.26 points, and produces Level 3 reports with higher factual consistency and analytical depth than commercial deep research systems, validated by both multi-LLM judging and human expert evaluation, showing a robust alignment (Pearson r=0.551).</abstract>
<identifier type="citekey">zhao-etal-2026-advancing</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.621/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>14256</start>
<end>14290</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Advancing ESG Intelligence: An Expert-level Agent and Comprehensive Benchmark for Sustainable Finance
%A Zhao, Yilei
%A Zhang, Wentao
%A Xiao, Lei
%A Zheng, Yandan
%A Liu, Mengpu
%A Lim, Wei Yang Bryan
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F zhao-etal-2026-advancing
%X Environmental, social, and governance (ESG) criteria are essential for evaluating corporate sustainability, yet professional ESG analysis is hindered by data fragmentation and complex multi-step workflows. Existing ESG benchmarks rely on LLM-generated questions that often suffer from semantic ambiguity, limiting their diagnostic reliability. We introduce ESGAgent, a hierarchical multi-agent system with a domain-oriented toolset for automating end-to-end ESG workflows, and a three-level benchmark derived from 310 corporate sustainability reports, featuring expert-sourced questions with verified, unambiguous answers at Levels 1-2 and full report generation at Level 3. Evaluations across 20 state-of-the-art closed-source LLMs show that ESGAgent achieves 84.15% accuracy on Levels 1-2, surpassing the strongest baseline by 3.26 points, and produces Level 3 reports with higher factual consistency and analytical depth than commercial deep research systems, validated by both multi-LLM judging and human expert evaluation, showing a robust alignment (Pearson r=0.551).
%U https://aclanthology.org/2026.emnlp-main.621/
%P 14256-14290
Markdown (Informal)
[Advancing ESG Intelligence: An Expert-level Agent and Comprehensive Benchmark for Sustainable Finance](https://aclanthology.org/2026.emnlp-main.621/) (Zhao et al., EMNLP 2026)
ACL