top of page

A Case Study on High-Impact AI Maths Tutoring from Third Space Learning

​

​Executive Summary

This case study provides an independent evaluation of Skye, Third Space Learning’s high-impact AI maths tutor. Third Space Learning (TSL) works with UK primary and secondary schools and US elementary and middle schools, providing annual licenses that allow schools to offer every child spoken, high-impact AI tutoring programmes with Skye, their AI tutor.

​

Following a rapid evidence review, we explored the extent to which Skye is aligned with core evidence principles for high-impact tutoring. In addition, we examined the initial evidence of the impact of Skye on teaching and learning outcomes, and highlighted the evidence of impact, the potential implementation challenges, as well as further opportunities and best practices.

​

Skye demonstrates strong alignment with core research principles for effective, high-impact AI tutoring. Its "I do, we do, you do" scaffolded structure aligns with step-based learning, breaking complex problems into manageable components. It delivers immediate, targeted feedback through probes, hints, and correction, prioritises guidance over answer generation, and ensures teachers maintain agency over topic selection.

​

Our evaluation reveals promising initial outcomes across multiple dimensions. Students demonstrated substantial within-session learning gains, improving from 34% accuracy on diagnostic check-in questions to 92% on check-out assessments. Confidence metrics showed similar positive trajectories, with 59.4% of students ending sessions with higher confidence levels and 66% showing session-on-session confidence increases over time. Two learner groups emerged as particular beneficiaries: anxious or quiet pupils who hesitate to seek help in traditional classroom settings found the non-judgmental AI interaction more accessible, and students requiring immediate consolidation after teaching benefited from the targeted, adaptive practice.

 

Skye demonstrates strong alignment with ethical AI tutoring principles through its curriculum-appropriate content, user agency preservation, and age-appropriate dialogue. The system maintains human-centred learning with teachers retaining professional judgment over topic selection and student participation. However, challenges remain, particularly with regards to inclusion such as in the case of Additional Language (EAL) learners and students with speech recognition difficulties who sometimes find it difficult to communicate with Skye.

A Case Study of High-Impact AI Maths Tutoring (1).png

Keywords

​

Maths, high-impact, AI, Tutoring, Evaluation, Case Study, Third Space Learning, Skye

bottom of page