To understand how caregivers and older adults might respond to AI-enabled technologies, we organized a co-design workshop to allow these prospective users to imagine and visualize AI solutions for themselves. The purpose of this design process was to surface user preferences and attitudes about AI products and services for caregivers – in advance of their commercial exposure to such solutions. The MIT AgeLab recruited informal family caregivers, adults ages 65 and older currently receiving care, and adults ages 75 and older to participate in a research workshop; these participants were split into smaller workshop groups. The goal for each workshop group was to come up with an imagined prototype of an AI solution for caregiving. Most of the workshop groups designed their prototypes around one of two main challenges: managing a care recipient’s behavior with reminders, recommendations, and scheduling assistance, and care coordination between family members and medical professionals. These findings underscore the results from previous research that care coordination and behavioral management constitute significant unmet needs for caregivers. In their designs, caregivers and older adults were willing to imagine expansive, “do-it-all” solutions. They tended to identify multiple challenges that they wanted to address, and accordingly dreamed up tools that would have multiple uses. Their multifaceted ideas may reflect the uncertain but highly touted potential of AI – a sense that the technology might be capable of doing many things very effectively and all at the same time, as well as the complexity of caregiving itself.