AI-powered applications are increasingly used in knowledge-intensive expert work, raising important questions about their trustworthiness, ethics, and reliability. However, little is known about how professionals actually develop trust in these systems in their daily work.

Based on interviews with nine accounting professionals, this study shows that trust in intelligent automation is shaped by a combination of technical, social, and contextual factors. Participants emphasized the importance of participatory development practices, including opportunities to contribute to system design and maintain open communication with developers.

The findings also highlight the active role of domain experts as technology users: professionals rely on their situational expertise and professional judgment when evaluating the reliability and usefulness of AI systems. As a result, trust in intelligent automation emerges across multiple levels—through direct interaction with AI systems as well as interpersonal collaboration and communication.

Overall, the study contributes practical insights into how trust in AI develops in expert work environments, emphasizing the importance of human-centered development and collaborative design.