Study goals
To investigate whether ChatGPT is capable of reliably and accurately extracting, calculating, comparing, and interpreting financial indicators of companies in the energy sector in a way that is useful for financial decision-making.
Relevance / originality
The evolution of AI models based on Natural Language Processing (NLP) is a prominent contemporary topic that should be analyzed from the perspective of accounting sciences.
Methodology / approach
Prompts were developed to extract, calculate, and compare financial indicators of three companies in the energy sector, covering three fiscal years, and the results were compared with the authors’ manual calculations.
Main results
ChatGPT was able to accurately present most of the requested indicators, as well as provide interpretations consistent with the specialized literature; however, there were issues in interpreting terms not explicitly stated in the financial statements analyzed, highlighting the need for technically clear
Theoretical / methodological contributions
It reinforces the importance of mastering prompt formulation and having a deep understanding of the technical fundamentals of financial indicators when using ChatGPT for the extraction, calculation, and interpretation of accounting indicators.
Social / management contributions
ChatGPT can democratize access to agile technical interpretations when properly instructed; however, the model does not replace professional judgment, especially in the analysis of non-recurring events, qualitative interpretations, and broader strategic considerations.