AI-Powered Climate Risk Communication

We deliver an AI-powered, multilingual, multi-hazard early warning service that provides clear, localised alerts to underserved, high-risk communities at scale.

The problem we solve

Climate forecasts are available, but they are not understood, trusted, or actionable.

Forecasts are often too technical, not available in local languages, or delivered through inaccessible channels. As a result, communities cannot act in time.

Resurgence uses DARAJA + AI to make climate early warnings accessible at scale, especially in low-resource languages and low-bandwidth settings. We translate official meteorological forecasts into clear, localised audio and text alerts delivered through trusted community channels in real time.

Aligned with UN Early Warnings for All (Pillars 3 and 4), our human-reviewed system extends – not replaces – existing national systems, ensuring life-saving climate information reaches everyone, everywhere.

Our impact

DARAJA + AI is being deployed in a conflict-affected context, the country of Sudan, to deliver multilingual early warnings in Sudanese Arabic and Tigre – reaching high-risk populations through radio, voice, and mobile channels.

What we deliver

DARAJA + AI: Early Warning for All

A first AI-powered, multilingual, multi-hazard, community-scale early warning system, built on 10 years of early warning delivery and community trust.

Multilingual, Real-Time Conversion

We transform official meteorological forecasts into clear, actionable text and audio alerts in multiple local languages simultaneously and in real time, enabling rapid comprehension across diverse populations.

Low-Bandwidth, Multi-Channel Reach

Our system is designed for fragile and low-connectivity environments, delivering warnings via SMS, Interactive Voice Response (IVR), radio, and platforms such as WhatsApp and Telegram – ensuring accessibility even on basic phones without internet access.

Human-in-the-Loop Governance

Every alert is validated through trained local forecasters, and community-informed safeguards before distribution, ensuring accuracy, cultural relevance, ethical oversight, and sustained trust.

Systems Integration, Capacity Building & Open Source

We embed directly within national meteorological agencies and strengthen long-term capacity through forecaster training, while using open-source, non-proprietary architectures to ensure sustainability, interoperability, and full partner ownership.

Our Partnerships under DARAJA + AI

How it works

Phase 1

Forecast Ingestion

We ingest and process official meteorological forecasts in real time from our national weather service partners.

Phase 2

AI Translation & Localisation

Our multilingual climate context models convert forecast data into clear, culturally adapted text in local languages.

Phase 3

Voice & Distribution

We generate accessible audio warnings and distribute them via WhatsApp, radio, SMS, IVR voice calls, and trusted community networks.

Phase 4

Continuous Learning & Scaling

Our human-in-the-loop review validates every warning before it reaches communities. Their feedback continuously improves our accuracy, reach, and trust. Ensuring we can scale across high-risk geographies, including fragile contexts such as Sudan.

Build systems that save lives

Partner with us to design inclusive early warning systems that reach the people who need them most.