Novissi Togo: ML and phone data to deliver shock-responsive social protection

Novissi Togo: ML and phone data to deliver shock-responsive social protection

Togo flag

Togo

Public Sector

High replicability and adaptation

Implementing Organisation

Ministry of Digital Economy and Digital Transformation, Government of Togo

Togo, Togo

Government

Implementing Point of Contact

World Bank Group

Contributor of the Impact Story

World Bank Group

Year of implementation

2020

Problem statement

The COVID-19 crisis disrupted livelihoods in Togo, particularly for informal sector workers who lack social protection coverage. Traditional social safety systems were not shock-responsive, lacked dynamic registries, and relied on outdated data and in-person processes, limiting the ability to rapidly identify and serve vulnerable households during the crisis. The NOVISSI program in Togo uses artificial intelligence to make social protection shock-responsive. Launched during the COVID-19 crisis, NOVISSI combines administrative data with satellite imagery and data science methods to assess needs and prioritize the most vulnerable informal workers for cash transfers. Machine learning models help determine eligibility and target beneficiaries more effectively in the absence of up-to-date social registries, enabling rapid, phone-enabled registration, assessment, and benefit delivery. This AI-enhanced approach supports scalable, data-driven targeting of social assistance to those most affected by economic shocks.

Submission Overview

The Government of Togo launched the Novissi program in 2020 as a fully digital social protection initiative in response to the socioeconomic disruption caused by COVID-19. The program was implemented under the leadership of the Ministry of Digital Economy and Digital Transformation, in partnership with national statistical authorities and international development partners. Novissi was designed to deliver rapid, targeted cash transfers to vulnerable populations affected by lockdown measures. Leveraging Togo’s national voter ID database and mobile money infrastructure, the government deployed machine learning and geospatial poverty targeting techniques to identify eligible beneficiaries, particularly in rural and economically distressed regions. The initiative represents one of the first large-scale applications of AI-driven satellite imagery and poverty mapping for social protection targeting in Sub-Saharan Africa. By integrating mobile money, digital identity systems, and remote sensing analytics, Novissi demonstrated how digital public infrastructure can enable rapid, transparent, and inclusive government-to-person (G2P) transfers.

AI Technology Used

Machine Learning

Key Outcomes

Efficiency & Productivity

Economic Value Creation

Access & Reach

Inclusion & Equity

Resource Efficiency

Accuracy & Quality Improvement

Novissi is a fully digital, AI-enabled social protection program launched by the Government of Togo to deliver rapid cash transfers to vulnerable populations during COVID-19 by combining satellite-based poverty mapping, machine learning targeting, digital ID verification, and mobile money disbursement. In the absence of a national social registry, the system used geospatial analytics to identify the poorest cantons and enabled eligible citizens to self-register via USSD using voter IDs, eliminating in-person processes. Within months, the program reached over 920,000 beneficiaries nationwide, prioritizing informal workers and women, who received higher transfer amounts. By leveraging existing digital public infrastructure, Novissi reduced administrative delays, minimized leakage risks, and enabled transparent, traceable government-to-person payments. The model demonstrated that AI-driven targeting can rapidly expand social protection coverage at national scale in low-income settings, while remaining deployable in resource-constrained environments.

Impact Metrics

Proportion of female beneficiaries

Baseline Value

Traditional safety nets did not systematically prioritize women in informal sectors Percentage of total beneficiaries

Post-Implementation

Women received higher transfer amounts and represented a majority share of beneficiaries Percentage of total beneficiaries

Jan 2020 - Dec 2020

Total beneficiaries reached through AI-targeted digital cash transfers

Baseline Value

NA beneficiaries under AI-enabled digital social protection system

Post-Implementation

Over 920,000 individuals received cash transfers Individuals (Number of beneficiaries)

Independent Evaluation·Apr 2020 - Dec 2021

Share of beneficiaries located in the poorest regions identified through machine learning poverty mapping

Baseline Value

No AI-driven geographic targeting system was in place Geographic coverage (Number of cantons targeted)

Post-Implementation

Program expanded from initial urban targeting to nationwide poverty-ranked cantons using satellite ML models Geographic coverage (Number of cantons targeted)

Independent Evaluation·Jan 2020 - Jan 2021

Implementation Context

Scaled

Togo

Informal sector workers, low-income and poor households, rural populations, and women

Key Partnerships

Ministry of Digital Economy & Transformation, Government of Togo, World Bank Group, mobile network operators, and academic and technical collaborators including some referenced researchers

Replicability & Adaptation

High

1. Mobile technology and digital payment infrastructure 2. Administrative, geospatial, and mobile operator data sources 3. Machine learning 4. Technical capacity in terms of development and operationalization 5. Government and partner coordination mechanisms

Supporting Materials

* The data presented is self-reported by the respective organisations. Readers should consult the original sources for further details.