Research
Solving for motivation and understanding at scale in K-12 students require a dual approach: facilitating human interactions to develop genuine motivation to learn and leveraging AI to deepen understanding.
Our research teams explore ways to unlock new human-AI synergies in education.
Building long-term motivation
Unlocking the motivation to learn is fundamentally a human endeavour. We research how to best facilitate the teacher, peer, and parent relationships that drive learning using the right technology.
- Cohort-based live classes
- Practice
- SmartFest
Developing AI models to detect misconception
Moving beyond the “right answer.” We develop AI optimized for the learning process itself, helping students actively unpack and resolve their own misconceptions.
- Practice
- AI Tutor model development
- Datasets
Enhancing content and pedagogy
How do we turn aggregate data on student misconceptions into targeted interventions? By iterating Math and Science instruction around the hurdles students face most. In the age of AI, this proactive approach helps students build the foundational mastery required to solve real-world problems.
- Cohort-based live classes
- Practice
Scaling open-ended assessments
Generative AI allows us to finally move past multiple-choice testing at scale. We are advancing open-ended assessments as the standard—fostering essential productive struggle for students and delivering actionable learning insights to teachers.
- Practice
- AI Tutor model development
- Datasets
>5 million students. >30 terabytes of insights.
Meeting the rigorous standards of formal education
Scaling evidence-based strategies
We start with the replicable evidence, then build scalable technology to broaden student impact.
A/B testing & random controlled trials
We validate every innovation across our geographically diverse student cohorts, ensuring statistical proof before classroom deployment.
Ground-level validation
We meet students where they live and learn. By engaging with them across homes, schools, and communities, we uncover the human realities behind the numbers.
Built for the masses
We experiment with solutions that are low-cost to implement so that they have the potential to help all students worldwide.