This paper explores the role of AI in supporting decision-making processes at construction sites, with a focus on scaffolding safety. Site managers must be able to balance safety, logistics, quality assurance, and risk management under critical time constraints. This study examines/explores how AI can improve decision-making in construction projects using scaffolding as an empirical case study. Scaffolding is a temporary structure critical for construction, and site managers must ensure it remains safe by identifying irregularities, confirming compliance with regulations, and taking corrective actions. The decision-making process for ongoing supervision of scaffolding involves three steps: identifying changes, verifying compliance, and acting, all of which require both practical experience and knowledge of regulations. The study is carried out partly through observations at a demonstration of AI applications for decision support, partly through semi-structured interviews, with on-site management personnel from construction companies. The study indicates that AI can improve decision-making by processing large amounts of information quickly, and hence improving decision-making related to cognitive limitations, time constraints, and information availability. However, to trust AI-tools, as a help and support in decision making, the AI needs to prove itself reliable as the site manager will still be the decision maker.