Artificial Intelligence in the analysis of public contracts and the oversight of public works

Abstract

This study analyzes the application of Artificial Intelligence techniques to support the preliminary assessment of public procurement procedures, administrative contracts, and public works oversight, with emphasis on the fields of civil engineering and public administration. The relevance of this topic stems from the role of Courts of Accounts in external control, particularly in the preventive assessment of the legality, legitimacy, cost-effectiveness, efficiency, and compliance of public contracting. Although recent advances in these technologies have expanded the possibilities for automated information analysis, systematic studies addressing their use in regulatory institutions remain limited. To address this gap, the study conducted a systematic literature review on approaches based on machine learning, natural language processing, and large language models. The results reveal consistent applications of these techniques in the automated interpretation of bidding documents and contracts, the detection of anomalous patterns in quantitative variables, cost estimation, risk identification, and the prediction of contractual events, such as amendments and disputes. A convergence between text analysis methods and predictive modeling was also observed, expanding the capacity to process both structured and unstructured information. It is concluded that the incorporation of these technologies into institutional control environments can strengthen external control by improving analytical efficiency, reducing operational bottlenecks, expanding the scope of assessments, and contributing to risk mitigation and the prevention of irregularities in public contracting.