Innovation Across Borders: Sri Lanka
Kavindya (Kavi) Thennakoon, co-founder of Tilli, reflects on what it takes to help teachers bring play, behavioural science, and accessible artificial intelligence (AI) tools into classrooms
Across 190 countries and territories, UNICEF colleagues and partners are on the frontlines of the greatest challenges affecting the lives of children and young people. Innovation Across Borders highlights the experiences, successes and learnings of innovation champions, committed to making positive social impact.
How did Tilli’s relationship with the UNICEF Venture Fund begin, and how has that partnership shaped your growth as a company?
Back in 2020, my co-founder Vidya and I came across a call for proposals from the UNICEF Venture Fund focused on digital solutions for child safety. After a very rigorous due diligence process, they became our first institutional investor and the first to take a chance on us. They were with us from our very first pilot, reaching around 50 children in Sri Lanka, to supporting Tilli’s growth to more than 30,000 children across four countries.
What have been your biggest challenges, high points, and lessons learned?
Our biggest challenge has been finding the sweet spot between three things: delivering high-quality learning that leads to measurable change for children; working within the realities of teachers’ packed schedules and high burnout rates; and remaining operationally and financially sustainable.
One value we hold very deeply is that no school should be unable to use Tilli because of cost. We’re also obsessed with capturing behavioural data and feedback at every touchpoint. From the beginning, we’ve built with teachers, not for them, and that’s been one of our biggest strengths.
What led to shifting Tilli’s towards a data- and AI-driven model that works through teachers?
When we started, Tilli was a child-facing, play-based social-emotional learning app, and we later began integrating more opportunities for co-play between children and caregivers. As I dove deeper into research on screen time — and was pushed by my advisors at Stanford to think critically about how we use screens alongside what it takes to create systems-level change in early childhood education — it became very clear that we needed to focus on classrooms and start by supporting teachers. There was also a major gap in high-quality data on children’s social-emotional and cognitive development, especially in countries like ours.
Today, we train teachers and education specialists to deliver social-emotional learning, and children participate in hands-on, play-based sessions for about 20-30 minutes each week over a 36-week curriculum, with a strong focus on building emotional awareness and regulation.
How are you using technology to strengthen learning and wellbeing without relying on direct digital exposure for children?
Screens come into play primarily for measurement and insight. Teachers and caregivers complete behavioural checklists for each child at three points in time, observing behaviour both at school and at home. Children also complete short, direct assessments.
This is where we use AI to help educators and school leaders engage with their student data conversationally. We’ve built an intuitive chat interface where educators can ask things like, “Can you summarize the last behavioural assessment and highlight two areas I should focus on this week?” or “Based on my class’s emotional regulation data, suggest an intervention for the next three weeks.” It enables personalized, culturally specific, data-driven teaching and learning at scale.
Is there one message you’d share with founders and innovators globally?
I’d say this: believe in the power of small, scrappy experiments. Learn from the people who know best — the teachers and children themselves — and hold onto the idea that everything we build should always remain a prototype.
