Long-term dementia care is often shaped by fragmented documentation practices that fail to capture the relational and evolving nature of caregiving. We present CareInsights, an AI-enabled infrastructure that supports caregivers by passively capturing interaction data, enriching it with contextual and clinical signals, and generating narrative-based feedback to support day-to-day person-centered care. Our system was developed through a two-year collaboration with seven long-term care facilities and integrated into existing electronic health record systems and memory therapy platforms to minimize disruption and streamline uptake. We evaluated CareInsights through a combination of system performance analysis, content assessments, caregiver surveys, and interviews. Our model demonstrated expert-level performance, and caregivers reported outcomes consistent with more effective person-centered dementia care. Adoption was supported by integration into existing staff workflows and design choices that mirrored caregiver reasoning. We conclude with design implications for "narrative AI infrastructure," offering guidance for scaling person-centered, AI-enabled support in real-world care environments. Please see the attached PDF for more information.

Accepted publication:

Dylan Edward Moore, Songyun Tao, Sophia R. R. Moore, Brian Morgan, Dio Tadin, Christina Sapp Tadin, and Elizabeth L. Murnane. 2025. CareInsights: AI-Enabled Infrastructure for Person-Centered Dementia Care in Resource-Constrained Facilities. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 9, 4, Article 202 (December 2025), 40 pages. https://doi.org/10. 1145/3770687

Memcara Inc.