Behind the Screen: Teachers Co-Designing WhatsApp AI Tools

As AI becomes an increasingly prominent part of classrooms around the world, questions about its role in teaching and learning are intensifying. Yet too often, teachers are missing from the conversation. At the heart of our three Teacher-in-the-Lead sandboxes lies a simple hypothesis: If teachers are meaningfully engaged in shaping how AI is used, we will unlock more learning. 

Aga Khan Foundation, Rising Academies, and Taleemabad are running AI sandboxes with teachers to try to reduce workload, support lesson planning, and enhance the quality of instruction. But crucially, teachers aren’t just end-users in this process—they’re co-designers.

Teachers-in-the-Lead Sandboxes

Sandbox teams are exploring what happens when teachers lead the design of AI tools for education from the beginning. The teams spend time in classrooms to understand the current challenges, observing how teachers currently use existing tools, and continuously iterating based on their feedback. This isn’t a one-off consultation; teacher expertise is woven into design, testing, and refinement at every stage of the project. The question driving this work is: how can teacher-led design and implementation ground AI tools in classroom realities, ensuring they’re shaped by educators’ expertise rather than technical possibilities alone? This approach is guiding our collective learning agenda and helping the broader sector to understand not just what AI can do, but how teachers can meaningfully drive development in the age of AI.

Meet Fidele and Mahrah, co-designing AI tools on WhatsApp in Rwanda and Pakistan

Here we want to share early insights from two of our sandbox teams who are co-designing a WhatsApp-based chatbot for teachers: Rising Academies in Rwanda and Taleemabad in Pakistan. These insights come from a recent conversation held in early December 2025.

Watch How Tari Works 👇

Watch How Lesson Genie Works 👇

Why WhatsApp?

When we asked Fidele and Mahrah why they chose WhatsApp as the platform for their AI tools, the answer was simple: it’s what teachers already use.

“Every teacher has WhatsApp on their phone,” Fidele explained. “We didn’t want to create something that required new devices or a heavy internet. WhatsApp is accessible, familiar, and works even in rural areas with low data.”

Mahrah echoed this, highlighting the practical realities in multi-grade classrooms: “Teachers are juggling two, three, sometimes five grades in one room. They don’t have time for complex apps. WhatsApp lets them get lesson plans instantly, in a format they already know how to use.”

For both Rising Academies and Taleemabad, WhatsApp wasn’t just convenient—it was the bridge that allowed teacher-designed support to reach the classrooms where it’s needed most, without disrupting existing workflows.

How have teachers driven the design of your AI tool?

Both Fidele and Mahrah emphasised that teachers are not just end-users, they are co-designers.

“We spent time in the field, interacting with teachers and identifying their ‘Jobs to Be Done’,” Mahrah explained. “We were actually performing the tasks teachers do, like planning lessons or juggling multiple grades to understand their challenges. Teachers also showed us how they currently use tools, like WhatsApp or ChatGPT, to get ideas for activities or lesson topics. This helped us design a solution that truly supports them in the classroom.”

Fidele described a similar approach, sharing, “We selected schools and held initial sessions with teachers, then created WhatsApp groups to engage them. We’re ready to keep learning from teachers and co-designing a number of things on the app together with them.”

For both teams, teacher input isn’t a one-off consultation, it’s woven into the design, testing, and iteration process. Teacher experience and expertise drives tech decisions for building the chatbot. 

What language does your chatbot support and why?

Both Rising Academies and Taleemabad have made design decisions shaped by linguistic realities. In each context, educators may teach in a language which differs from their mother tongue. Rising Academies’ chatbot Tari, now operates entirely in English. Earlier versions supported Kinyarwanda, but the Ministry requested the team focus fully on English to help teachers navigate the ongoing transition from French to English as the official medium of instruction. 

Fidele, “What the Ministry wants is to make sure that we are supporting English as a medium of instruction. We know teachers sometimes struggle with pronunciation, so you can type a sentence, and Tari will read it out in English.”

Mahrah shared a different perspective as Taleemabad’s Lesson Genie is designed to respond in a bilingual mix of English and Urdu. “Teachers can message the bot in Urdu, and it replies in both languages to make lesson plans easy to follow. We chose this hybrid approach because many teachers find fully English materials difficult to use in practice, and the goal is to support, not overwhelm them as they manage demanding multi-grade classrooms.”

What are you most excited about?

Mahrah responded, “For us, the most exciting part is giving teachers instant, practical lesson support, which they already use every day. Many of our teachers are in multi-grade classrooms, sometimes teaching two, three, or even five grades in one room. The Lesson Genie can generate grade-wise lesson plans within seconds, adapted to the amount of time a teacher actually has… 20 minutes, 30 minutes, whatever the reality is.”

“We also built in a feedback loop. Teachers can rate plans, suggest improvements, and tell us what’s working,” Mahrah continued, “that’s really important because it means the tool keeps evolving with them.”

Fidele shared, “I’m excited by the image-to-lesson-plan feature. A teacher can just snap a picture of a guide or a page from a book, and the bot instantly summarises it or turns it into a usable lesson plan. That saves so much time. There is also pronunciation support. Since we’ve transitioned from French, many teachers struggle with English pronunciation. Now they can type any sentence, and WhatsApp immediately models it for them.”

Early previews from Fidele and Mahrah of chatbots developed in Rwanda and Pakistan show what’s possible when teachers lead the design of AI tools for education. But this is just the beginning.

Over the next couple of months, we’re excited to share ongoing insights from a series of in-person events with each Teacher-in-the-Lead sandbox, bringing together lead teachers, head teachers, and education officials. 

If you’re working across AI in education or simply curious about what teacher-led design of AI implementation in classrooms looks like in practice, we’ll be sharing more as the Teacher-in-the-Lead sandboxes continue to evolve.

---

EdTech Hub’s AI Observatory and Action Lab is made possible by support from UK International Development. Please follow along and join the conversation on LinkedIn!