Person-centered dementia care depends on caregivers’ ability to interpret subtle behavioral changes and share relational knowledge across shifts, roles, and families. However, high staff turnover, fragmented documentation, and inconsistent family engagement often erode continuity of care in long-term, resource-constrained care facilities.

This project introduces an AI-enabled narrative infrastructure designed to overcome these issues and support the full caregiving sensemaking loop: 

• Capturing meaningful moments during care interactions
• Enriching them with clinical and contextual information
• Synthesizing patterns over time
• Delivering actionable, role-specific insight at the point of care

Developed through a multi-year collaboration with seven long-term care facilities in the Northeastern U.S. and Canada, the platform integrates directly into existing workflows, including memory-therapy sessions and electronic documentation systems. Rather than introducing a separate dashboard, it captures bounded care interactions, enriches them with contextual signals, and generates narrative summaries aligned with caregiver reasoning. 

In real-world deployments, caregivers reported faster access to resident context, reduced repeated mismatched interventions, improved situational awareness across shifts, and greater confidence in care decisions while families reported increased transparency, improved trust, and more productive care conferences.

Overall, this work advances a model of narrative AI infrastructure: systems designed not only to document care but to preserve and circulate relational knowledge in complex, high-turnover environments.

Memory care specialists at Memcara: Christina Tadin, Dio Tadin, Brian Morgan