ABSTRACT
Artificial intelligence (AI) is increasingly used as a source of emotional support, reflection, and perceived companionship. Despite growing public and clinical interest in AI companionship, little research has examined how individuals construct these experiences within broader relational systems. Drawing on systemic, attachment, and social constructionist perspectives, this study examines AI companionship using anonymized data from Psyhelp, an AI-assisted mental health platform. A qualitative-dominant mixed methods design integrated thematic analysis of user narratives with engagement patterns and self-reported emotional indicators. Findings suggest that users primarily experienced AI companionship as a relational supplement characterized by availability, responsiveness, and psychological safety. Four themes emerged: AI as a relational supplement, a transitional emotional regulator, a psychologically safe disclosure space, and a facilitator of human relationships. Findings suggest that AI companionship reflects broader relational needs and has implications for systemic assessment, case conceptualization, and clinical training.