AI-driven disaster response platform enhances task quality by 21.5 points through structured RAG and adaptive routing.

Why It Matters

This research highlights how human-centered AI design can overcome critical 'last-mile' challenges in disaster management. By transforming fragmented data into actionable plans and optimizing resource allocation, such systems can lead to more effective and timely interventions, ultimately saving lives and resources.

Key Finding

The AI platform significantly boosted the quality of disaster response tasks, optimized resource routing to reduce computational costs while maintaining near-optimal response times, and provided reliable offline guidance on mobile devices.

Key Findings

• Improved overall task-quality scores from 61.4 to 82.9 (+21.5 points) compared to a standard RAG baseline.
• Reduced solver calls by up to 85% while remaining within 7–12% of optimal response time.
• Delivered fully offline mobile guidance with sub-500 ms response latency and 54 tokens/s throughput on commodity smartphones.

Research Evidence

Aim: How can an AI-powered, human-centered platform effectively bridge the gap between fragmented disaster data and coordinated field actions to improve response quality and efficiency?

Method: Platform development and simulation-based evaluation

Procedure: Developed ResQConnect, an AI platform featuring specialized agents for data extraction and task planning using a structured RAG workflow, an adaptive event-triggered multi-commodity routing algorithm, and a compressed language model for offline mobile guidance. Evaluated performance through realistic flood and landslide scenarios.

Context: Disaster management and response in hazard-prone regions

Design Principle

Human-centered AI systems for critical operations must balance data processing accuracy, operational efficiency, and resilience through adaptive algorithms and offline capabilities.

How to Apply

When designing systems for emergency response or other time-sensitive, data-intensive fields, consider incorporating AI agents for data interpretation, adaptive routing algorithms for dynamic resource allocation, and offline functionalities for guaranteed operation.

Limitations

Performance was evaluated through simulations; real-world deployment may encounter unforeseen variables. The effectiveness of the compressed language model may vary across different smartphone hardware.

Student Guide (IB Design Technology)

Simple Explanation: This study shows that using smart AI tools can make disaster response much better by helping people make better decisions faster, even when communication is difficult.

Why This Matters: This research demonstrates how advanced AI can be applied to solve real-world problems, improving human safety and efficiency in critical situations, which is a key aspect of user-centered design.

Critical Thinking: To what extent can AI fully replace human judgment in high-stakes disaster response, and what are the ethical considerations of relying on AI for life-or-death decisions?

IA-Ready Paragraph: The development of ResQConnect demonstrates the significant impact of human-centered AI on disaster response. By employing a structured Retrieval-Augmented Generation (RAG) workflow and an adaptive event-triggered routing algorithm, the platform achieved a substantial improvement in task quality (from 61.4 to 82.9) and operational efficiency, while also providing crucial offline mobile guidance. This approach highlights the potential for AI to overcome critical 'last-mile' challenges in high-risk environments, ensuring more effective and resilient operations.

Project Tips

• Consider how AI can help users process complex information in your design project.
• Explore adaptive algorithms for systems that need to respond to changing conditions.
• Think about offline functionality for any design that might be used in areas with poor connectivity.

How to Use in IA

• Reference this study when discussing the use of AI for data analysis and decision support in your design project.
• Cite the findings on improved task quality and efficiency when justifying your design choices for a complex system.

Examiner Tips

• When evaluating a design project, look for evidence of how the system addresses 'last-mile' challenges and ensures user effectiveness.
• Assess whether the design considers resilience and offline capabilities for critical applications.

Independent Variable: AI-powered platform features (structured RAG, adaptive routing, offline model)

Dependent Variable: Task quality scores, response time, solver call reduction, mobile guidance latency and throughput

Controlled Variables: Realistic flood and landslide scenarios, standard RAG baseline, commodity smartphones

Strengths

• Addresses a critical real-world problem with significant societal impact.
• Integrates multiple advanced AI techniques (RAG, multi-agent systems, adaptive routing, compressed models).
• Evaluated using realistic scenarios and quantitative metrics.

Critical Questions

• How scalable is this platform to different types of disasters or geographical regions?
• What are the potential failure modes of the AI agents, and how are they mitigated?
• What is the user training requirement for effectively operating this AI-assisted system?

Extended Essay Application

• Investigate the impact of different AI agent specializations on disaster response effectiveness.
• Explore the trade-offs between route optimization algorithms and computational resources in disaster scenarios.
• Develop and test a simplified offline AI guidance system for a specific emergency situation.

Source

ResQConnect: An AI-Powered Multi-Agentic Platform for Human-Centered and Resilient Disaster Response · Sustainability · 2026 · 10.3390/su18021014