This study aimed to develop a model for the adoption of artificial intelligence (AI) in financial risk management across the gold value chain. The study was designed using a constructivist grounded theory approach with an interpretivist orientation and was aligned with the “Research Onion” logic to harmonize philosophy, strategy, and method. Data were collected through in-depth semi-structured interviews with purposively and then theoretically sampled experts from different segments of the gold value chain; data collection and analysis proceeded concurrently. Analysis employed initial coding, focused coding, and theorizing, alongside constant comparison and analytic memoing. Trustworthiness and dependability were enhanced through member checking, peer debriefing, and inter-coder agreement. The findings revealed a core process of human–machine hybrid financial decision-making, shaped by the dynamic interplay of three dimensions: building trust and the gradual adoption of AI; data-driven governance and infrastructure; and financial risks and the strategies for managing them. In this configuration, AI functioned as a decision-support tool while humans remained the final arbiter; trust in models was contingent on data quality, explainability, and contextual fit, and implementation outcomes reflexively recalibrated trust levels and governance arrangements. The contribution lies in proposing a context-sensitive process model for the responsible integration of AI into financial risk management in the gold industry and in explicating the mechanisms of trust, governance, and human–machine integration, yielding practical implications for data governance design, workforce capability building, and the development of explainable models.