ResQConnect: AI-assisted disaster response and task synthesis system with agentic RAG

Disaster response operations demand rapid, coordinated decision-making under degraded infrastructure and limited connectivity. Existing systems rely on manual workflows and fragmented information, leaving coordinators ill-equipped to process requests efficiently. ResQConnect is a human-centered AI-assisted decision-support platform integrating four components: an Agentic RAG workflow that converts unstructured help requests into field-executable task plans; an AET multi-commodity routing engine for dynamic resource allocation; a fine-tuned lightweight language model enabling offline guidance on mobile devices; and an RL-based retrieval policy that minimizes unnecessary cost. Human coordinators retain full decision authority throughout. The system contributes to SDG 11 by enhancing resilience in climate-related emergencies.