The Influence of Reliable Model Operations on Trust and The Adoption of AI Systems

ModelOps, MLOps, AI trust, Artificial Intelligence Adoption, Machine Learning Operations, trustworthy AI, Model Governance

As Artificial Intelligence (AI) systems become increasingly integrated into business operations and daily life, the challenge of building trustworthy AI has emerged as a critical determinant of successful adoption. While significant attention was given to algorithmic innovation, the operational frameworks that govern AI model lifecycle management, collectively known as Model Operations (ModelOps), remain an underexplored factor in shaping stakeholder trust in AI systems. The study of these operational frameworks becomes even more critical in the era of generative and agentic AI. This study investigates how various ModelOps practices shape trust in AI systems through a quantitative survey of 266 AI professionals, business stakeholders and end users across diverse industries. Using validated measurement instruments, we examine how different dimensions of ModelOps influence trust in AI systems. Our findings reveal that organisations with reliable model operations are associated with higher levels of stakeholder trust in AI systems (β = 0.46, p < 0.002). In turn, trust significantly predicts both perceived value (β = 0.44, p < 0.001) and actual AI usage (β = 0.91, p < 0.001). Notably, trust was found to fully mediate the relationship between operational maturity and adoption outcomes, suggesting that technical reliability alone is insufficient to drive usage without first establishing a foundation of trust. These findings establish that in the era of generative and agentic AI, investing in robust ModelOps practices, specifically monitoring, quality, and human-in-the-loop oversight, is as crucial as algorithmic refinement for realising sustained organisational impact.