From risk management to AI-enabled governance: five intellectual orientations in EPC and infrastructure project risk governance

Purpose This study aims to synthesise the intellectual development of engineering–procurement–construction and infrastructure project risk governance and examine whether bibliometric evidence supports five overlapping intellectual orientations. Design/methodology/approach The study combined bibliometric analysis with Bibliometrix/Biblioshiny, VOSviewer and a scripted workflow to examine 173 Scopus-indexed publications (1999–2026). The analytical corpus was derived through relevance validation of 678 candidate records, with provenance checked against an 801-row master ledger. Fifteen provisional 2026 records were excluded from temporal comparisons. Findings The analysis identified five orientations: risk analysis and control (53 documents), performance and overrun (80), governance and relational mechanisms (52), digital information integration (7) and artificial intelligence (AI)-enabled decision support (18). The share of period publications addressing performance and overrun increased from 34.7 % in 1999–2018 to 59.7 % in 2023–2025, whereas the digital and AI orientations remained comparatively limited. These patterns concern the published literature and do not establish organizational maturity or realised capability. Research limitations/implications Scopus coverage, title/abstract-based relevance validation, citation lag, analytical thresholds and interpretive coding delimit the findings. Project-level research is required to test the proposed relationships. Practical implications The orientations provide structured review prompts for risk processes, outcome linkage, decision rights, information integration and human oversight of predictive systems; they are not intended as a diagnostic maturity instrument. Originality/value The study connects construction risk, megaproject governance and construction AI through an auditable corpus and an evidence-bounded conceptual framework.