Case study · EdTech
Self Innovation
Custom AI tutor agents adapted to each learner's pace, in three languages.
- Client
- Self Innovation
- Year
- 2024
- Category
- EdTech
Stack
Challenge
What we were solving.
Self Innovation runs vocational training programs across rural India with learners across English, Hindi, and Telugu. Their tutors couldn't personalize to every learner's pace, and the wide language mix made one-size content ineffective.
They wanted an AI tutor that adapts to each learner's pace, language preference, and prior performance — with safe-handoff to humans on tough material.
Solution
How we built it.
We built a multi-modal AI tutor on Claude that delivers lessons via text + audio + image, adapts difficulty based on quiz performance, and gracefully escalates to human mentors when confidence drops.
Content is authored once in English and translated + voiced through a curated translation pipeline. Mentors monitor learning sessions through a real-time dashboard.
Results
The numbers that mattered.
- Engagement
- 2×
- Languages
- 3
- Completion lift
- +47%
- Mentor time saved
- 60%
Related work
More from the portfolio.
Get started
Want a result like this?
Tell us about your project — we'll come back with a fixed scope and timeline within 48 hours.