AI or human: How the type of information to be disclosed alters customer service agent preferences

While prior research has largely examined service providers' recommendation strategies and consumer responses, less is known about how consumers form preferences in the initial stages of interactive recommendation—when disclosing personal information to service agents. This research examines how the type of information disclosed (subjective vs. objective) influences consumers’ preferences for AI versus human customer service agents. Across four studies, we find a systematic matching effect: consumers prefer AI agents when disclosing objective information but human agents when disclosing subjective information. Study 3 identifies trust in recommendation capability as the underlying mechanism, while Study 4 shows that AI anthropomorphism moderates this effect. Beyond advancing theory, our findings reveal that aligning information type with agent type enhances both consumer satisfaction and purchase intentions. Together, this work contributes to the literature on information disclosure, human–AI interaction, and service design.