How locus of causality shapes human-AI decision-making

Purpose – Despite the recent surge in studies on implementing Artificial Intelligence (AI) in different B2C settings, extant work on human-AI interaction in the B2B context are predominantly conceptual. To addressthis gap, the present work provides empirical evidence on the factors shaping human-AI interaction in business service interactions. Design/methodology/approach – In two vignette-based experimental studies and one field experiment, we examine the role of a businessinteraction’s essential but overlooked outcome (successful vs. unsuccessful)in the attributional process, answering whether managers’ locus of causality toward human (vs. AI) company representatives mediates satisfaction from the interaction. Data were collected from marketing managers with B2B experience. They were randomly assigned to scenarios describing a business conversation between a human and another company representative, human or AI agent. Findings – Our study demonstrates that company managers attributed AI success to external factors in the case of successful performance. In contrast, the success of human agents was attributed internally. Moreover, the external attribution of successful AI diminishes anticipated satisfaction from working with the AI. Originality/value – Our findings deepen our understanding of the psychological mechanisms that shape human-AI interaction and provide actionable insights for integrating AI into the business service. They contribute to discussions on attribution processestowards AI,studies on satisfaction with the use of AI, and tests of the mediating nature of locus of causality in shaping managers’ satisfaction with AI collaboration.