AI for Foundational Learning: Applied Lessons from Teacher and Student Support Tools in India

• In Event: AI-Enhanced Adaptive Teaching: From Evidence to Implementation

Proposal

Over 260 million children are taught by 8 million teachers across 1.5 million schools in India. Despite an enrollment rate of 95%, over half of Indian children are unable to read simple texts by age 10. A shortfall of over a million teachers, combined with weak instructional support, further exacerbates learning gaps — particularly for students from disadvantaged backgrounds.
At the Central Square Foundation (CSF), we believe that technology has the potential to democratise quality education, and have seen evidence of its efficacy across foundational and middle grades, both globally and in India. We are now exploring how AI can better support teachers and students, particularly in low-resource settings. In this panel contribution, we reflect on two ongoing pilots designed to address core challenges in teaching and learning. Rather than present EdTech as a silver bullet, we aim to discuss what it takes to meaningfully deploy useful AI-powered solutions that are pedagogically aligned and contextually relevant.
The first innovation is an AI Teacher Coach — an on-demand coaching tool that provides teachers personalized feedback based on analysis of audio recordings of their classroom interactions. Grounded in NIPUN's (India's FLN mission) pedagogical principles, the tool provides curriculum-aligned lesson plans and evaluates classroom delivery on parameters such as student engagement and instructional alignment. It offers personalised, and non-evaluative feedback to teachers on demand, with local language and offline functionality for low-connectivity areas. While still in early stages, the pilot tests not only the tool's technical accuracy, but also evaluates user experience, training quality, and teacher perception of AI. Furthermore, it provides insights on the tool's technical accuracy, usability, and feedback quality, alongside how teachers interpret and act on the tool's insights.
The second innovation — an AI-powered Personal Tutor — is designed to provide individualized learning support to students, thereby replicating the benefits of one-on-one tutoring at scale. Delivered through conversational AI, the tutor encourages dialogue and guides the student step-by-step, building conceptual clarity. The pilot uncovers what it takes to contextualise the product for low-income school segments in India and encourage meaningful engagement. Furthermore, it explores the effectiveness of training, support mechanisms, and the role played by teachers and parents in facilitating meaningful usage.
Together, these innovations explore important considerations regarding the meaningful design, deployment and augmentation of teaching-learning through AI in education. These pilots are not just viewed as interventions, but as learning experiments — to understand how AI might be deployed for equity. Their approach is grounded in a broader theory: that strengthening teaching with safe and intelligent tools — rooted in evidence, inclusivity and context, and guided by pedagogy — can build bridges between evidence, policy, and practice. In doing so, this work contributes to emerging Global South perspectives on the responsible use of AI in pedagogy, and aims to engage with re-examining inclusion and social equity through educational transformation.

Author

• Gouri Gupta, Central Square Foundation