Pilot Launch | Powering Anticipatory Action for Floods in Senegal using AI and Smart Contracts
Flood in the city center of Dakar, Senegal - Image licensed from Shutterstock
Mercy Corps Ventures’ Anticipatory Action Accelerator received over 230 applications from more than 50 countries worldwide. Emerging from that competitive pool, we are excited to share the AI-powered, end-to-end solution that will protect communities vulnerable to climate shocks in the Thiès region of Senegal.
This pioneering initiative is OPAL for Floods, a consortium led by Data-Pop Alliance (DPA), together with Wasdi, Jokalante, and Lamport Lab. Together, they're building a localized platform that harnesses Artificial Intelligence, satellite imagery, and smart contracts to shift flood response from reactive to proactive—enabling communities and authorities to act before disasters strike.
This post is the first in a two-part series documenting this pilot. The second blog will share key insights and results after the pilot is completed and evaluated.
In Brief
The pilot will develop and deploy an AI-driven flood prediction and early action system that integrates satellite data, predictive analytics, and automated cash transfers to enable faster, more effective responses.
The pilot will protect ecosystems and improve resilience for climate-vulnerable communities in Grand Dakar and Thiès, Senegal—and lays the groundwork for expansion to neighboring regions in Senegal and West Africa.
A key innovation is the integration of smart contracts to automate emergency fund disbursement, reducing response times from days to hours while ensuring transparency and accountability.
If successful, this model could scale to other flood-prone regions globally, setting a new standard for AI-powered anticipatory action in humanitarian response.
Image licensed from Shutterstock
The Problem
Floods are increasing in frequency and impact
Floods are becoming more frequent, intense, and destructive—posing a growing threat to ecosystems, human health, and the global climate. For some years now, Senegal has been suffering from disruption to its rainfall pattern, which has a significant impact on a large part of its communities living in risk areas. Despite the efforts made by the authorities each year in flood control, the problem persists. The August 2022 floods caused by 3 days of heavy rainfall displaced thousands of people destroying homes and critical infrastructure in Grand Dakar and Thièt. In 2024, torrential and late rains caused flooding that displaced more than 56,000 people and damaged many health facilities and schools. These events demonstrate the urgent need for improved risk monitoring and anticipatory action tools and mechanisms.
Globally, disasters like floods contribute to economic damages exceeding $140 billion annually and an estimated 300,000 premature deaths. In recent years, West Africa has become a climate hotspot, with flood alerts increasing dramatically due to extreme weather patterns and inadequate early warning systems.
We know when floods start—but the response is often too slow
While early detection technologies exist, critical gaps remain in translating data into action. Current challenges include:
Fragmented data systems: Flood-related information is siloed, inconsistent, or too technical for non-expert users to leverage effectively.
Delayed response cycles: Manual verification of hazard triggers, beneficiary selection, and fund disbursement can take days or weeks, reducing the effectiveness of assistance.
Lack of localized insights: National-level models fail to capture sub-national variations, limiting the ability to prioritize interventions at the community level where they're needed most.
Limited AI integration: Most tools are designed externally without adaptation to local languages, decision-making structures, or institutional capacities.
Even with high-resolution data, the ability to predict how floods will evolve and trigger proactive interventions remains limited. This gap forces communities into reactive modes that increase losses of lives, livelihoods, and infrastructure.
The Pilot
Mercy Corps Ventures is partnering with the OPAL for Floods consortium to bridge the gap between cutting-edge data science and local action. The pilot will deploy an integrated platform in Grand Dakar and Thiès (ADM3 level), covering approximately 24,000 hectares where the 2022 floods demonstrated urgent needs.
What makes OPAL for Floods innovative
The platform integrates multiple technologies to create a seamless anticipatory action workflow:
1. AI-Powered flood analytics
Real-time satellite imagery and Earth Observation data to detect and predict flood risks
Geospatial intelligence tools to visualize flood impact, population exposure, and infrastructure vulnerability
A natural language chatbot assistant that allows non-technical users to query the system in local languages
2. Smart contract automation
The platform will use smart contracts to automate critical, time-intensive steps:
Trigger validation: Automatically verify when parametric flood thresholds (water level, rainfall intensity) are met using satellite data
Beneficiary selection: Cross-reference pre-established vulnerability criteria to identify affected households
Automated disbursement: Execute 10-100 mobile money transfers per flood event to pre-identified recipients within hours instead of days
By automating these workflows, smart contracts drastically reduce administrative costs, accelerate aid delivery, and ensure transparency through immutable execution logs.
3. Community-centred communication
Early warning messages translated into local languages (Wolof, Peul) and delivered via SMS, voice calls, and WhatsApp
Two-way communication channels through Jokalante's network of 200,000+ users
Context-appropriate messaging based on population demographics
Image courtesy of Mercy Corps
Who Will Benefit
The pilot will directly support:
Over 50 practitioners from local and national government agencies (including the National Agency of Civil Aviation and Meteorology, Ministry of Hydrology)
100,000+ residents in flood-exposed neighborhoods through better-targeted interventions and rapid financial support via smart contracts
Local partners who will redistribute funds to 500-2,000 beneficiaries per event
Beyond immediate beneficiaries, the pilot will strengthen national capacity to integrate AI and automation into disaster risk management, creating a replicable model for anticipatory action.
Learning agenda
We will test three core hypotheses during this pilot across a multiple indicators:
Building the evidence base for Anticipatory Action
The selection of OPAL for Floods marks more than just the conclusion of our first Anticipatory Action Accelerator—it represents the beginning of a new chapter in how we approach humanitarian challenges. By supporting solutions that combine cutting-edge technology with deep community engagement and financial innovation, we're helping to build a future where vulnerable populations are protected before crises strike, not just supported after disaster hits.
Finally, we have recently conducted a research study analysing the results of the Accelerator’s call in the wider context of the Anticipatory Action innovation ecosystem. We will be sharing these findings soon – stay tuned! Here are some of the insights from the Accelerator’s applications:
233 applications received from 57 countries
Most applications came from Eastern (40%) and Western Africa (16%)
18% came from Kenya, 9% from Nigeria and Uganda respectively
Flooding was the most addressed climate hazard (44%), followed by drought (41%), heatwaves (7%), cyclones (5%), storms (2%), and wildfires (2%)
66% of all proposals target either smallholder farmers (34%) or low-income populations (32%)
82% of proposals use emerging technologies as a core of the solution
26% use a combination of two or more emerging technologies
96% use Artificial Intelligence / Machine Learning
93% use them to improve forecasts or predictive analysis
38% of proposed solutions were in the concept design phase, 20% completed Proof-of-Concept, 17% had a Minimum Viable Product (MVP), and 15% already validated their solution being ready for deployment.
To learn more about the vision and methodology behind the Anticipatory Action Accelerator, read our foundational piece here.