The contribution of AI-based stakeholder network modeling to risk governance in housing development

Purpose This paper aims to examine how artificial intelligence (AI) can enhance risk governance in housing development by integrating stakeholder-centric network modeling, predictive machine learning and adaptive reinforcement learning (RL). Design/methodology/approach A hybrid framework is developed that integrates stakeholder theory with AI techniques. Network analytics identify central stakeholders and contractual bottlenecks; supervised models forecast risk trajectories; and an RL agent adapts contractual and scheduling decisions. The framework is evaluated on 50 synthetic housing projects calibrated against real-world benchmarks. Findings Compared with a rule-based baseline, the AI-driven framework reduced financial risk by 20% (95% CI ±1.5%, n = 50), lowered legal disputes by 15% (±1.2%) and improved environmental, social and governance compliance by 30% (±2.0%) across the simulated housing portfolio. These gains demonstrate that AI can complement traditional methods by offering dynamic, data-driven insights into multi-stakeholder coordination and governance. Practical implications The framework provides actionable support for project managers and regulators by improving early risk detection, enhancing inter-stakeholder alignment and accelerating compliance with evolving sustainability standards. Originality/value This study advances stakeholder theory by embedding AI-based analytics into governance processes for housing development. It contributes a replicable model that bridges theoretical perspectives with operational decision-making, showing how AI can augment rather than replace conventional expertise in real estate risk management.