Sensebod: Offline, Local-Language AI Tutoring with Examination Toolkit for Equitable Learning
Added 2 Mar 2026
About this innovation
Innovation Details
Executive Summary
Sensebod (from local Krio, meaning ?a smart student?) is an offline-first, local-language AI tutor with a built-in examination toolkit. It delivers personalised tutoring, mentorship, gamified engagement, and exam preparation for secondary school students, fully aligned with BECE/WASSCE curricula. Unlike global platforms such as UNICEF?s Learning Passport, Sensebod is built for African realities: it runs offline on EasySTEM Servers, supports learning in local language and adapts to the exam systems students face. Piloted with 400 students across 25 schools, Sensebod was co-developed with students, teachers, and parents to ensure cultural relevance and curriculum alignment. Ministries of Education and Communication anchor policy alignment, UNICEF supported deployment through the DLH in a Bag, and Africell is enabling zero-rating and offline data syncing to guarantee nationwide access. Early indicators are promising: 65% of students are weekly active users averaging one hour per session, and parents report 32% more study time. These trends suggest potential gains of 0.20?0.35 standard deviations in comprehension and exam readiness within one academic year equivalent to several additional months of learning. Teachers benefit from dashboards that flag at-risk learners early, while policymakers gain real-time, disaggregated data to guide equity-focused resource allocation. The societal and economic impact is twofold: improved exam performance and equity for marginalised learners, and reduced remediation costs while building a prepared workforce. By 2027, Sensebod will be open-sourced as a certified Digital Public Good, and by 2035 it will reach six million learners across Africa a replicable, ethical AI model for inclusive, exam-ready education.
Background
Sierra Leone?s education system faces a crisis of access, quality, and equity. Pass rates for the West African Senior School Certificate Examination (WASSCE) pass rates remain persistently low, with many students entering secondary school lacking the foundations to succeed. Overcrowding compounds the challenge: average class sizes exceed 50 pupils, and schools face an estimated deficit of two classrooms each (ERIC, 2023). In addition, Teachers are overstretched, particularly in rural districts where multi-grade teaching without adequate materials is common (Education Above All, 2025). Language barriers worsen inequity: while most instruction is in English, many students think and communicate in Krio, Temne, or Mende. This mismatch fuels disengagement and dropout, disproportionately affecting girls and children with disabilities (MoE, 2021). Families with means turn to private tutors, but most cannot afford them, leaving the system tilted toward privilege. Digital learning tools often fail in Sierra Leone?s context. Nearly 70% of the population remains offline (DataReportal, 2024), yet most platforms assume constant internet. In addition, content is largely English-only and rarely aligned with Basic Education Certificate Exam (BECE)/WASSCE curricula, excluding learners who think in local languages. These gaps highlight the need for an offline, inclusive, curriculum-aligned AI solution built for African realities. To address these challenges, In 2023, EasySTEM, with UNICEF and the Government of Sierra Leone, piloted the EasySTEM Server hosting Sensebod v1, giving secondary students offline access to digital learning even in remote areas. The Digital Learning Hub in a Bag (DLH) built on this by providing in extension solar power, and devices to cover connectivity, content, and capacity. Pilots showed two lessons: offline infrastructure works, but static content is insufficient. Students, teachers, and parents demanded personalised learning, interactive practice, and exam-prep support. Sensebod v2 was developed directly from these user insights through a user-centred design process. Its development was funded by the African Centre for Technology Studies (ACTS) via the National Science, Technology and Innovation Council (NSTIC), with offline deployment support from UNICEF and the Peace Corps, and business acceleration from the GoTogether Accelerator in London and training infrastructure for models via Amazon web services partnership. Sensebod acts as a one-on-one AI tutor, adapting to each learner?s pace and providing explanations in English and local languages. Its examination toolkit includes adaptive quizzes, flashcards, and revision plans aligned with BECE/WASSCE, available both online and offline through the EasySTEM Server. Teacher dashboards flag struggling learners early, while parents will soon receive simple weekly progress updates. Running on solar-powered EasySTEM Servers and DLH kits, Sensebod remains resilient to connectivity and electricity challenges that derail most global platforms. Sensebod operationalises Sierra Leone?s National Digital Learning Strategy by bridging the digital divide through offline delivery (Pillar 1), strengthening student and teacher capacity (Pillar 2), and localising curriculum content (Pillar 3). It also advances the Radical Inclusion Policy (2021) by embedding equity for girls, rural learners, and students with disabilities through accessible design. Regionally, it supports the AU Continental Education Strategy for Africa (CESA 16?25).
Strong partnerships reinforce these ambitions. UNICEF supported offline deployment via the easystem server. Peace Corps extended rural reach through volunteers. Africell is enabling zero-rating for affordability, offline-to-online syncing via towers, and gamification rewards such as airtime linked to learning streaks. AWS awarded $70,000 in cloud credits to accelerate model development. Endorsements from the Vice President and the Ministers of Education and ICT underscore Sensebod?s standing as a national priority.
Sensebod?s significance reaches beyond education. Socially, it dismantles barriers of geography, gender, language, and disability, ensuring equal opportunities for learners nationwide. Economically, it lowers tutoring costs, builds local AI capacity, and prepares a future-ready workforce. Sensebod is built on open-source tools, reflecting EasySTEM?s commitment to open innovation. By 2026, it will be certified as a UN-endorsed Digital Public Good (DPG), ensuring affordability, preventing proprietary lock-in, and enabling replication across ECOWAS. This approach strengthens open ecosystems, accelerates localisation, and positions Sensebod as a globally shareable model of responsible EdTech.
Sensebod is built on top of a host of open source tools.Hence, with EasySTEM?s commitment to open innovation, the roadmap is to certify Sensebod as a Digital Public Good (DPG) by 2026, aligning with the UN-endorsed DPG Standard. This makes it not only a national tool but also a globally shareable innovation, one that strengthens open ecosystems and accelerates localisation, ensuring the avoidance of proprietary lock-in and replication across Africa. By 2035, Sensebod aims to reach six million learners across Africa, representing a generational investment in human capital, bridging policy ambition and classroom reality while scaling responsibly, inclusively, and sustainably.
Problem Statement
Sierra Leone?s education system is in crisis. Regardless of recent strides, close to 40% of students fail or do not meet university entry requirements in the WASSCE (MBSSE, 2021), with many reaching secondary school lacking core foundations. Overcrowding worsens learning: class sizes average 50+, and over 60% of teachers report unmanageable workloads (ERIC, 2023). Evidence confirms overcrowding is among the strongest predictors of poor comprehension and disengagement in low-income countries (UNESCO GEM Report, 2021). In rural areas, shortages of trained teachers and materials compound the problem (Education Above All, 2025). Language further deepens inequity: though research shows children learn best in their first language (Cummins, 2000; Ouane & Glanz, 2010), most lessons are delivered in English, driving dropout and weak literacy, particularly for girls and learners with disabilities (MoE, 2021). Digital learning could, in theory, close these gaps, but in practice, most solutions fail. Nearly 70% of Sierra Leoneans remain offline (Data Reportal, 2024), yet global platforms assume connectivity, deliver in English only, and ignore BECE/WASSCE curricula. Evidence from Sub-Saharan Africa shows imported platforms without localisation rarely scale (Trucano, 2020). Sierra Leone, therefore, faces a critical gap: it needs offline-ready, curriculum-aligned, and inclusive AI solutions that bridge policy ambition and classroom reality. Without this, millions, especially girls, children with disabilities, and rural learners, risk exclusion. Economically, the country cannot afford a generation unprepared for work; building human capital is central to the National Development Plan and Africa?s Agenda 2063. AI uniquely addresses these challenges by enabling personalised support at scale, in local languages, and offline, capabilities absent in current solutions.
Objectives
Primary Objective:
Improve learning outcomes for Sierra Leonean and West African students through AI-driven, curriculum-aligned tutoring, mentorship, and gamification, all accessible via an examination preparation toolkit. Secondary Objectives:
- Expand equitable access by ensuring offline delivery, local language access, and curriculum contextualisation.
- Support government priorities by aligning with the National Digital Learning Strategy, Radical Inclusion Policy, and Digital Development Strategy.
- Generate rigorous evidence on comprehension, exam performance, and engagement to strengthen AI in education policymaking.
- Build a scalable regional model for inclusive AI, adaptable across ECOWAS countries. Cross-Cutting Gender & Inclusion Objectives:
- Ensure gender equality, disability inclusion, and equitable access by embedding local language in education, accessibility features, and inclusive design from the outset.
- Strengthen participation of girls, rural learners, and students with disabilities, while empowering women educators and caregivers as key users. Goals Guiding the Development of Sensebod:
Development is grounded in user-centred design, localisation to curricula and languages, resilience for offline and solar-powered environments, and sustainability through partnerships with telecoms, ministries, and universities.
Full submission details are available to the owner and platform administrators.
Log in