WHA- RealTime Reporting AI-SyS

System Overview

WHA–RealTime Reporting is an AI-driven system developed to support the World Health Assembly (WHA), held annually at the United Nations in Geneva, Switzerland. The platform assists the WHO team in streamlining and automating the process of summarizing high-level speeches delivered by Member States and delegates during WHA live sessions, a task that was previously performed by over 80 staff members.

During proceedings, the system captures official statements from Member States and delegates in real time, extracts key policy priorities and thematic elements, and generates concise, structured summaries tailored for the Director-General and senior leadership. It integrates advanced natural language processing techniques and leverages a WHO-trained model built on an extensive repository of past WHA speeches, datasets, and validated summaries to ensure accuracy, consistency, and adherence to institutional style guidelines.

In addition to that, the platform provides exportable structured data outputs and an advanced interactive dashboard for the WHO team. The system also employs an iterative model refinement pipeline and leverages a curated corpus of generated summaries as part of its continuous refinement process, enabling adaptive learning of WHO-specific linguistic nuances, domain terminology, and stylistic conventions. By integrating newly ingested speech transcripts, WHO-provided datasets, and structured feedback mechanisms, it continuously fine-tunes performance, enhances contextual accuracy, and strengthens alignment with institutional reporting standards over time, ensuring long-term knowledge retention and sustained performance enhancement.