Abstract

New York City operates the largest municipal homelessness system in the United States, providing nightly shelter to more than 85,000 individuals, including over 17,000 families with children. Despite substantial public investments, outreach operations remain fragmented across agencies, resulting in delays in care, inconsistent follow-through, and repeated crisis encounters. National studies indicate that individuals experiencing homelessness miss 50–60% of medical and social-service appointments, and outreach teams spend significant time on manual coordination activities. These inefficiencies contribute to recurring emergency-room utilization and preventable health and safety crises. This work explores HOCI - Homeless Outreach Coordination Intelligence an agent-based AI platform designed to enhance real-time coordination among vulnerable individuals, outreach workers, and service providers. Leveraging multilingual conversational interfaces, voice-enabled encounter logging, trauma-informed user experience design, and geospatial intelligence, HOCI serves as an intelligent coordination layer, linking individuals to shelters, food resources, healthcare, mental health support, and urgent crisis services. The platform supports case managers and program directors through automated service matching, appointment scheduling, and citywide analytics

Download

Share

COinS