Machine Learning-Enabled Construction Project Management: Systematic Review, Comparative Performance Synthesis and Implementation Framework

PurposeMachine learning (ML), a subset of artificial intelligence (AI), is increasingly transforming construction project management (CPM) processes, yet a significant gap exists between research advancements and practical implementation. This study synthesizes empirical findings and proposes a framework to bridge the research-practice divide. Design/methodology/approachA systematic review was conducted using Scopus and Web of Science databases, following PRISMA guidelines. 64 peer-reviewed studies published between 2015 and 2025 were selected and coded using open and axial methods. The analysis was guided by systems theory, complexity theory, and dynamic capabilities theory to develop the Machine Learning–Enabled Construction Project Management (MLCPM) framework. FindingsML use in CPM aligns with three core areas: (1) planning tasks like cost estimation and scheduling, (2) execution tasks such as progress tracking and risk detection and (3) monitoring activities including cost control, delay prediction and performance tracking. The MLCPM framework offers a four-layer architecture consisting of data, ML processing, integration and output, with feedback and retraining loops to support scalable deployment. Furthermore, a quantitative synthesis of ten comparative studies revealed that hybrid and ensemble methods achieved superior performance in cost estimation (cost) (e.g. LightGBM and DNN-SVR), tree-based ensembles (random forest (RF) and gradient boosted trees (GBTs)) showed optimal accuracy for duration forecasting (duration) and hybrids with metaheuristic optimization outperformed single algorithms in delay prediction (delay). Originality/valueThis review consolidates existing literature on ML applications in CPM and introduces a comprehensive MLCPM Framework. Additionally, a quantitative subset synthesis of comparative evaluations extracts task–method performance patterns, enhancing decision relevance. We provide practical, theory-grounded guidance for adopting ML in construction.