LLM-powered question answering for Environmental, Social, and Governance (ESG) reports
| dc.contributor.author | Tran, Vinh-Khanh | en |
| dc.contributor.author | Tran, Khanh-Tung | en |
| dc.contributor.author | Pham, Quoc-Viet | en |
| dc.contributor.author | Guha, Krishnendu | en |
| dc.contributor.author | Nguyen, Hoang D. | en |
| dc.contributor.funder | Research Ireland | en |
| dc.contributor.funder | European Regional Development Fund | en |
| dc.date.accessioned | 2026-01-06T10:07:32Z | |
| dc.date.available | 2026-01-06T10:07:32Z | |
| dc.date.issued | 2025 | en |
| dc.description.abstract | Sustainability reporting has become a crucial tool for companies to demonstrate commitment and performance on Environmental, Social, and Governance (ESG) matters. However, implementing ESG reporting presents many challenges, including data standardization as well as conforming to various reporting frameworks and standards such as Global Reporting Initiative, Sustainability Accounting Standards Board, Task Force on Climate-related Financial Disclosures, and European Sustainability Reporting Board. These challenges are more prevalent, especially in small-to-medium enterprises. To address these issues, this paper proposes a question answering framework leveraging Retrieval-Augmented Generation combined with large language models to support ESG management in business environments. The framework aims to ease the required resources and efforts for ESG analysis, improve the accuracy and relevance of retrieved information, and aid organizations in analyzing their sustainability reports. Additionally, we explore different methods for handling heterogeneous data types in ESG reports, including text and tabular data retrieval strategies. Our implemented system, tested on data obtained from leading companies in Australia, showcases promising results in accommodating a wide variety of use cases in the ESG domain. This demonstrates the potential of the proposed framework as well as other Artificial Intelligence driven tools in assisting organizations in effectively managing ESG reporting and improving their sustainability performance. | en |
| dc.description.sponsorship | Research Ireland (12-RC-2289-P2) | en |
| dc.description.status | Peer reviewed | en |
| dc.description.version | Published Version | en |
| dc.format.mimetype | application/pdf | en |
| dc.identifier.citation | Tran, V.-K., Tran, K.-T., Pham, Q.-V., Guha, K. and Nguyen, H.D. (2025) 'LLM-powered question answering for Environmental, Social, and Governance (ESG) reports', AIQAM '25: Proceedings of the 2nd ACM Workshop in AI-powered Question & Answering Systems, pp. 49-53. https://doi.org/10.1145/3746274.3760395 | en |
| dc.identifier.doi | 10.1145/3746274.3760395 | en |
| dc.identifier.endpage | 53 | en |
| dc.identifier.isbn | 979-8-4007-2056-7 | en |
| dc.identifier.startpage | 49 | en |
| dc.identifier.uri | https://hdl.handle.net/10468/18364 | |
| dc.language.iso | en | en |
| dc.publisher | Association for Computing Machinery | en |
| dc.relation.ispartof | AIQAM'25 The 2nd ACM Workshop in AI-powered Question & Answering Systems in conjunction with ACM MM 2025, Dublin, Ireland, 27 October 2025 | en |
| dc.relation.project | info:eu-repo/grantAgreement/SFI/Centres for Research Training (CRT) Programme/18/CRT/6223/IE/SFI Centre for Research Training in Artificial Intelligence/ | en |
| dc.rights | © 2025, the Authors. This work is licensed under a Creative Commons Attribution-NoDerivatives 4.0 International License. | en |
| dc.rights.uri | https://creativecommons.org/licenses/by-nd/4.0/ | |
| dc.status | Not peer reviewed | en |
| dc.subject | Sustainability | en |
| dc.subject | ESG reporting | en |
| dc.subject | Artificial Intelligence | en |
| dc.subject | Retrieval-Augmented Generation | en |
| dc.subject | Heterogeneous document | en |
| dc.subject | Conversational | en |
| dc.title | LLM-powered question answering for Environmental, Social, and Governance (ESG) reports | en |
| dc.type | Conference item | en |
