RFI-to-Field Impact Analytics for Mega-Terminal Structural Packages: A Traceability Model Linking RFIs, Revisions, and Schedule Impacts

Requests for Information (RFIs) can delay and complicate large construction projects, yet their downstream effects on design revisions, field rework, and schedule are rarely traced. This paper proposes an RFI-to-field traceability framework for airport Project Management/Construction Management (PMCM) teams. The framework links each RFI to the resulting drawing revisions, field change orders (FCOs), Non-Conformance Reports (NCRs), rework, and critical-path delay across four project systems (Procore, Oracle Unifier, Primavera P6, and Bluebeam), and adds a leakage-aware classifier that assigns each RFI to a low, medium, or high impact class using only information available at submission. Because the logs of the motivating airport case are confidential, the framework is evaluated on a synthetic benchmark of 2,400 RFIs over 48 months with known ground-truth links; the results therefore validate the pipeline and are not empirical findings about real projects. The tiered linkage recovered 89%–99% of true links at each stage, with precision of 57%–96% (lowest for NCR and schedule links), and linkage-derived impact labels agreed with ground truth (Cohen's κ = 0.88). With a time-ordered split, the best early-warning model (random forest) reached a macro-F1 of 0.39 (95% CI 0.36–0.43) and ROC-AUC of 0.62 against a macro-F1 of 0.26 for a majority-class baseline, i.e., weak discrimination under the assumed noise. No cost or schedule saving is claimed; replication on anonymized real logs with a manual link audit is required.