Can an AI Teaching Assistant Improve Student Learning? Measuring the Impact of Shaia in Colombia

Header Block
In this Image A teacher and student in math class using the Shaia platform.

In many under-resourced settings, teachers lack access to the pedagogical guidance, instructional materials, and timely feedback that would help them teach better. AI-powered educational technologies offer a potential avenue for closing this gap. For instance, the education technology company MentuLabs developed Shaia, an AI-driven platform that supports teachers to reduce lesson preparation time and increase engagement with students. Specifically, Shaia supplies a specialized algorithm that assists instructors with their teaching, tracks their progress in the lesson plan during class, and provides after-class feedback. Can this type of program be an effective tool for teachers to strengthen their teaching practices and, in turn, improve student learning?

In collaboration with MentuLabs, in partnership with Shaia Foundation and the Education Secretariat of Bogotá, researchers are conducting a randomized evaluation to measure whether Shaia changes teaching practices and improves student learning. The intervention includes 90 schools (270 teachers and 6,000 students) across Bogotá. The schools have been randomly assigned to either receive the Shaia platform or serve as a comparison group.

Researchers will evaluate teachers’ attendance to Shaia training workshops, time spent on class preparation and feedback, teaching practices, and subjective assessments of students, while measuring students’ math test scores, technology use in class, and expectations to enroll in post-secondary and STEM programs.

Results will be available in late 2026.


Implementing Partners

MentuLabs
Shaia Foundation
Bogota Secretariat of Education

Funding Partner

Jacobs Foundation Logo