A digital learning platform leveraging AI to train community health workers in Senegal

Community health workers play a vital role in In Senegal as they are often the first point of contact for communities with medical needs living in rural areas a long way from hospital services. The work they do focuses on prevention, monitoring pregnant women and child health, teaching hygiene, and directing patients to healthcare centers when necessary. However, their training is often lacking: sessions are rare, cost a lot, and are not necessarily suitable for local circumstances. In addition, access to digital tools is still limited, particularly in areas that do not have a reliable Internet connection. These restrictions slow down sharing good practices and have a negative impact on the quality of services provided.

To take up these challenges, the kSanté platform is underpinned by the synergy of three levers:

• Offline accessibility: thanks to microSD cards, training content can be accessed without having an Internet connection, even in the most remote areas.
• Localized content: modules cover the main issues of maternal, child and community health, designed using guides approved by health authorities.
• An AI chatbot: it has been developed with PleIAs , a company specialized in training small open source AI models, and enables health workers to ask questions in real time and get personalized answers, still offline. The FID-financed project provides for evaluation through a randomized control trial involving 300 community health workers in Senegal: 200 trained using kSanté (modules and chatbot), and 100 following the traditional learning path. Impact on knowledge, practices and motivation will be measured after four months, through field surveys and analyzing usage data.

This pilot phase will enable data to be produced on kSanté's effectiveness and the conditions for implementing it. The main deliverables are:

• Reinforcing community health worker skills, better understanding of healthcare protocols and feeling more confident in their daily tasks;
• Tangible changes to services provided to communities, particularly in prevention and monitoring mothers and children;
• Information to guide public decision-makers about how much the solution is accepted and its level of value against investment;
• Bases for scaling-up, with the aim of training up to 100,000 community health workers by 2030 in several countries and languages. Over time, this innovation could contribute to reducing inequalities in access to healthcare and improving the health of rural communities in West Africa on a long-term basis.