Background: Conversational AI has been proposed as a potential form of support for older adults and family caregivers. However, less is known about how older adults and caregivers together negotiate AI’s role in later-life family care, including what forms of support are considered helpful, acceptable, or intrusive. Earlier prototype work and informal simulation-based evaluation suggested that functions such as contextual memory, emotional responsiveness, reminders, and caregiver summaries may be useful. Aims: To explore how older adults and caregivers understand and negotiate the potential role of conversational AI assistants in family care, and to examine the perceived utility and acceptability of prototype functions. Methods: This exploratory qualitative study forms Stage 1 of a PhD project and builds on previous prototype development by examining acceptability and early feasibility in family care settings. Interviews will be conducted with 10–15 older adults and caregivers in total. The interview guide explores views on autonomy, transparency, trust, responsibility, and control regarding AI assistants, as well as perceived usefulness and feedback on prototype functions. Results: Initial observations suggest that conversational AI may support reminders, emotional connection, and caregiver awareness, but its acceptability depends on trust, privacy, autonomy, and negotiated boundaries between support, monitoring, and control. Prototype functions may be useful, but require further refinement. Full findings will be presented. Implications: Findings will help designers and practitioners identify acceptable use cases, clarify boundaries of AI support, and develop context-sensitive approaches to conversational AI in later-life family care.