Generative Artificial intelligence in project formulation from a OPM perspective
Description
This study did not test a formal hypothesis in the traditional empirical sense; rather, it addressed four research questions concerning the integration of generative artificial intelligence (Gen-AI) into project formulation within the Organizational Project Management (OPM) framework. The underlying premise guiding the review was that, despite the rapid expansion of Gen-AI capabilities and their documented relevance to project management, the literature had not yet examined these capabilities as a distinct object of study within OPM-level formulation processes. The dataset comprises 91 documents cited throughout the manuscript. Of these, 56 constitute the core systematic review corpus, identified through a PRISMA-based search of Scopus and Web of Science as described in Section 4; the search returned 257 records, of which 163 were excluded at the identification stage and a further 38 at the eligibility stage. The remaining 35 documents support the conceptual framework (Section 2) and the synthesis of prior literature reviews on AI applications in OPM (Section 3). For each of the 56 documents in the systematic review corpus, the dataset records bibliographic information (authors, year, journal, country of affiliation), the OPM management domain addressed (project, program, portfolio, or a combination), the application sector, the project lifecycle stage examined, the AI technology or model reported, and the alignment with each of the four analytical axes (A1 to A4) defined in this study, corresponding to the four research questions. The data show that Gen-AI applications in project management and formulation are concentrated in isolated, sector-specific initiatives, predominantly in construction, with limited representation in manufacturing, finance, and healthcare. The most notable finding is the near-complete absence of Gen-AI from OPM-level formulation processes: of the three studies addressing Gen-AI within OPM, none extends beyond portfolio-level financial applications, leaving program and project management entirely unaddressed at this level. These patterns should be interpreted as indicative of the current state of a rapidly evolving field rather than as a definitive characterization, given that 80% of the corpus was published between 2023 and 2025. Researchers wishing to use this dataset can apply it to replicate the axis-by-axis classification, extend the search with updated date ranges, or use the coding scheme as a template for related reviews in adjacent technological domains. The dataset is available from the corresponding author upon reasonable request.
Files
Institutions
- Universidad Nacional de ColombiaBogota D.C., Bogotá