Students use an adaptive learning tool in their classroom during the project team’s supervision mission. Copyright: Way Production/World Bank
Access to education in Côte d’Ivoire has improved over the past decade, and enrollment rates continue to rise. However, learning outcomes remain low. A 2019 assessment shows only 17.2% of students reach sufficient proficiency levels in mathematics and 40.4% in reading.
More recently, the Service Delivery Indicators (SDI) 2022 report documented an average score of just 38.3 out of 100 in French and math combined; for example, 48% of grade 4 (CE2) students are unable to read a single word from a simple eight‑word sentence, and only 6.5% of students at this level can correctly multiply 37 times 42. Overly complex curricula and teaching practices that do not match students’ actual levels contribute to this gap.
But technology offers solutions. Through the World Bank’s Youth-RISE project, supported by the Mastercard Foundation, adaptive learning platforms are being introduced to help learners progress at their own pace. In Côte d’Ivoire, the pilot concerns around 2,000 students across 25 technical and vocational education and training (TVET) institutions, including 19 public schools.
This post takes you behind the scenes of this ambitious initiative, detailing its early results, implementation challenges, and future prospects within the Ivorian context.
What we know about adaptive learning
Adaptive learning refers to pedagogical systems that dynamically adjust content and teaching pace based on an individual learner’s demonstrated needs, often using artificial intelligence (AI) and learning algorithms. A growing body of international research shows that adaptive learning improves students’ academic performance and engagement. A recent systematic review of 69 studies found that 59% of the research reports a significant improvement in academic results, while 36% saw an increase in student engagement. Beyond academic gains, adaptative pedagogy is known to strengthen motivation and engagement. For instance, El-Sabagh (2021) showed that in an adaptive e-learning environment based on learning styles, students were more engaged than in a traditional e-learning context.
Positive impact in mathematics and French for active users
The project was implemented across 25 TVET institutions. In total, 1,529 learners, or 76.5% of the expected group, actively used the tool by completing at least one exercise. Specifically, 1,387 learners worked on French exercises and 1,343 on mathematics, over a 22-week period.
The impact analysis, conducted using the difference-in-differences method, revealed several key findings. In mathematics, the impact was strongly positive. Learners who actively used the tool saw their performance increase by 0.234 standard deviation (statistically significant at the 1% level), and simply being in a class with access to the tool led to a 0.230 standard deviation improvement (significant at the 5% level). Using learning‑time equivalence benchmarks, active use of the tool is associated with learning gains of approximately 11.2 months in mathematics. In addition, classroom‑level access to the tool, irrespective of individual usage, is linked to an average gain of about 11 months, likely reflecting spillover effects within classrooms rather than equivalence between passive exposure and active engagement.
In French, while being in a class equipped with the tool had no significant effect (-0.054), learners who actively used the tool improved by 0.121 standard deviation (significant at the 5% level). (Figure 1.) Accordingly, for French, learning gains are observed only among students who actively used the tool. Active engagement with the tool is associated with an increase equivalent to approximately 5.8 months of learning.
Figure 1: Adaptive learning in mathematics and French
Credit: Authors
A deeper analysis of progress showed that the tool was most beneficial for the learners who struggled initially. For example, in the French modules, the bottom 15% of students, based on initial test scores, demonstrated remarkable speed. They were between 6 to 15 times faster than the average-performing group and between 27 to 53 times faster than the highest-performing group. Indeed, for each performance group, relative learning speed is measured as the relative difference between the score obtained after a sufficiently large number of exercises, determined by the AI embedded in the tool, and the baseline learning assessment score.
These findings highlight how adaptive learning can significantly reduce educational inequalities by providing personalized support precisely where it is needed most.
Teachers and learners share their experience
The impact of adaptive learning is not just visible in the data; it’s felt directly in the classroom. Teachers have praised the platform for helping them revisit essential language fundamentals that are often difficult to address in traditional settings. As one teacher from the Technical and Vocational High School of Gagnoa put it: “The language improvement exercises served as a revival of pedagogical foundations. In our classrooms, it's often difficult to revisit certain language fundamentals, and these exercises have truly helped us to do so.”
Beyond language, educators reported a shift in students’ attitude toward mathematics. One observed that “Adaptiv’World sparks motivation and fosters a strong desire among learners to engage with mathematics”.
On the students’ side, the feedback is just as powerful. A learner shared how the tool transformed their relationship with the subject: “I didn’t understand mathematics before. But now, thanks to the platform, I actually want to study it.”
These voices from the field show that adaptive learning isn’t just a concept, it’s a real solution making a difference in how students learn and how teachers teach.
Perspectives for a successful scale-up of adaptive learning
The pilot has demonstrated the value of adaptative learning in Côte d’Ivoire. But there remain opportunities for improvement as we look toward scaling up in the future. In this context, several key policy questions need to be addressed:
- How can digital infrastructure be strengthened to ensure equitable access at scale?
- How can adaptive learning content be aligned with national curricula while preserving flexibility and effectiveness?
- What teacher training and support models are most effective for large‑scale implementation?
The work cited in this blog has benefitted from the financial support of the Mastercard Foundation. The World Bank values its partnership with the Foundation in advancing the agenda on building resilient education systems and digital skills for youth.
Acknowledgments: Waly Wane, Ali Coulibaly, Yves Jantzem, Modibo Sidibe, Juan Barón, Sophie Cerbelle, Anne Caroline Smith and Nguessan Enoh Ndri for their contributions; Ministry of Technical and Vocational Education and Training of Côte d’Ivoire (METFPA), including the General Directorate of Initial Training (DGFI), the General Inspectorate of Technical Education, Vocational Training and Apprenticeship (IGETFPA), and the Youth Employment and Skills Development Program (PEJEDEC); the Coordination Office for Employment Programs (BCP-Emploi); Mastercard Foundation, EvidenceB.
Join the Conversation