An AI-Assisted Multicriteria Analysis of Corporate Sustainability Reports_prompts2&3
Description
As academic advisor to Marina de Oliveira Guimarães (CNPq/INT scholarship holder, Brazil, 2024–2025) and Giulianna Costa Fico (FUNCATE scholarship holder, Brazil, same period), I, Adriane Cavalieri, based on conversations with both, declare that they developed projects formally documented in Technical Notes during the period mentioned above, which are held in the archives of the National Institute of Technology (INT, Brazil). Marina was responsible for developing the Generative Artificial Intelligence (GenAI) prompts, while Giulianna provided the (publicly available) corporate sustainability reports of the companies studied; both jointly analyzed the resulting data. These projects served as the basis for the development of another work, presented herein. It should be clarified, however, that the prompts contained in this database differ from those formally recorded in the aforementioned Technical Notes, and that the prompts developed by the scholarship holders during their fellowship period were not used in the construction of this database. Nonetheless, I wish to acknowledge and thank both scholarship holders for their research into the scientific literature and their dedication to the quantitative analysis, as well as CNPq and FUNCATE for the funding provided through their respective scholarships. The data underlying this work were derived from publicly available corporate sustainability reports. GenAI prompts were specifically developed to extract, translate, and analyze information from these reports, with data collection and analysis procedures embedded directly within the prompt design. The resulting analytical outputs included evidence logs, thematic matrices, analytical justifications, narrative summaries, and aggregated tables — consisting primarily of processed tabular data, which were subsequently translated and used to generate graphical representations in Microsoft Excel. All datasets and analytical outputs generated from Prompts 2 and 3 were systematically organized into structured folders and subfolders, as presented herein.
Files
Steps to reproduce
PROMPT 2: Aggregated Table of Sustainability Themes and Circular Practices for each selected company. PROMPT 3: Thematic_matrices ---->> MCDA_results ---->> MATRIX A ---->> MATRIX B --->> Circularity Index