IFRC x Microsoft

Agent Orchestration for Humanitarian Systems

AI-assisted humanitarian intelligence for faster, source-grounded crisis reporting. This site presents Team 21's final project report across requirements, engineering, testing, and handover documentation.

Project Abstract

Problem Statement

In a world where complex crises are constantly evolving, humanitarian teams spend excessive time collecting fragmented context from multiple sources, which delays evidence-based decision-making in a fast-moving situation. Teams of analysts, from both the IFRC and National Societies alike, are under immense pressure to provide vital insights within a short timeframe. This demand for speed creates an undesirable tradeoff between maintaining data integrity and providing structured information quickly: two elements that are essential in humanitarian decision-making.

Our Solution

Team 21 leverages agentic AI to develop the Virtual Situation Room (VSR), a comprehensive dashboard that delivers a wealth of information for humanitarian crises. The system follows a deterministic approach: it ingests data intelligently, generates automated reports, and displays validated reports on a clean user interface. The VSR maintains transparency by providing confidence scores and a bibliography, and supports reports with geospatial and geopolitical context, giving analysts the necessary information they require while minimizing cognitive load.

Achievement and Impact

The VSR reduces research and analysis from a couple of days to just a matter of hours. By centralizing the analyst workflow into a singular platform, it revolutionizes research, with analysts no longer needing to spend unnecessary time filtering through articles for relevant information. Furthermore, the Virtual Situation Room provides actionable summaries for operational managers looking for important context or recent updates. These various functionalities present a unique insight into what the future of crisis analysis may look like: user-friendly dashboards with comprehensive information, allowing analysts to focus on the tasks that matter.

Report Tabs and Evidence

Use these tabs to jump directly to detailed evidence pages.

Requirements and Research

Partner goals, personas, use cases, MoSCoW tables, alternatives, and technical decisions.

Requirements and Research

Algorithms and Design

Method pipeline, experiment setup, quantitative results, UI process, and system architecture.

UI Design and System Design

8-Minute Portfolio Video

Meet the Team

Ammar Khan

Ammar Khan (Team Lead)

Client Liaison, UI Design, Fullstack Engineer

  • Application Engineering
  • Frontend Design
  • Azure Integrations
  • Stakeholder Coordination
Asad Majeed

Muhammad Asad Majeed

AI Engineer and Platform Integration

  • Admin Workflow Engineering
  • Report Chatbot Integration
  • Core Agent Pipeline Research
Raahym Rehan

Raahym Rehan

AI and Data Engineer

  • AI Agent Orchestration
  • Data Ingestion Engineering
  • System Integration
Ismail Falih

Falih Latheef

Data Engineer, UI Design, Frontend Engineer

  • Frontend Design
  • Data Ingestion Engineering

Gantt Chart