Explainable Artificial Intelligence (XAI) model for transparent and trustworthy tender evaluation of construction projects

Construction tender evaluation is a high-stakes decision process in which contractor selection is expected to remain transparent and defensible. Although artificial intelligence (AI) effectively enhances analytical decision processing scalability using machine learning, AI adoption in project tender evaluation is constrained by limited interpretability and weak justification of AI insights. This study develops a conceptual Explainable Artificial Intelligence (XAI) tender evaluation model that integrates data preprocessing, predictive modeling, and SHAP explainability within three phases. The model provides decision insights at global and contractor levels through dataset-level feature attribution, contractor-level explanations of evaluation criteria and trade-offs, and project governance insights supporting audit trails and tender award justification. A pilot study was conducted among 10 Malaysian construction sector experts to examine the relevance and practical applicability of the proposed model. The findings indicate XAI strengthens for decision transparency, improves tender ranking interpretability, and supports transparent tender deliberation, whereas professional judgment remains central to a final tender decision award. This study strengthens the link between predictive analytics and procurement governance by explicitly revealing the interaction dynamics of ranking criteria that are often obscured in conventional tender evaluation. This study positions data governance as a prerequisite for credible explanations and decision support. Future research should empirically test the proposed model in live tender evaluation settings and establish sectoral standards for explainability and data governance for construction projects.