Teamwork in project management: Mapping AI as mediator with triangulated bibliometrics (1966–2023) under the IMOI framework

We map the intellectual structure of teamwork in project management (PM) using a triangulated bibliometric design to interrogate 16,320 Scopus records (1966–November 2023). Three complementary lenses—keyword co-occurrence (topical cohesion), bibliographic coupling (research front), and co-citation (conceptual foundations)—are interpreted through an IMOI team-effectiveness lens that organizes mediators (affective, behavioral, cognitive) by phase (forming, functioning, finishing). We also overlay AI-related terms and treat AI as a socio-technical mediator—a cognitive augmenter, a coordination infrastructure, and a learning/retention aid—rather than as a standalone topic. To enhance rigor, we document analytic choices in an Assumption & Parameter Log, apply anchor-based labeling rules (Interpretive Propositions), and assess robustness to threshold and resolution choices. The results reveal three consolidated domains—project success, agile methodologies, and knowledge management—and two emergent domains—virtual/distributed teams and leadership—concentrated in construction and software contexts, where ERP and product development are increasingly prevalent. Across the maps produced, three mediators recur: coordination under uncertainty (behavioral), shared mental models/transactive memory (cognitive), and trust/psychological safety (affective). AI is most visible at topical and research-front layers (decision support, BIM analytics, context-aware allocation) and less embedded in the field’s foundational canon. We contribute a mediator-first account of PM teamwork that clarifies where AI lands in team processes, offer phase-specific guidance for matching AI affordances to mediators, and demonstrate a reproducible mapping protocol for exploratory bibliometrics. Limitations (single database, concept-measure gaps for AI) suggest the desirability of phase-aware, mechanism-level studies and cross-sector comparisons.