Studied in a government school and college in Kerala, this is another story of empathy with intelligence of Ramamoorthy G, co-founder of Learning Matter, an EdTech Startup from Bangalore.
Driven by his personal experience of graduating from a government college, he recognised that finding enough expert human teachers to solve the critical English communication gap for 30 crore students is impossible. And so, he builds Tara AI, a use case to solve the education problem for students.
Tara AI is an autonomous voice technology teacher, utilising Narrow AI via an Android app or Alexa skill. Instead of just answering questions, Tara teaches lessons and asks students to respond, providing coded feedback on grammar and pronunciation.
This no-screen solution aligns with the Ministry of Education's NEP goals and has delivered 3.5 million hours of teaching to about 70,000 students. Moorthy advises fellow innovators to focus on AI applications that solve genuine problems, such as the shortage of teachers, rather than merely creating interfaces for existing tools. Here is the story of Mr Ramamoorthy building Tara AI.
Tell us something about yourself
Alright! Hello readers, I am Ramamoorthy G. For more than 10-15 years, I was deep in the technology field, working in software and tech companies. Later, I pivoted and spent many years in education, heading product and software development, even spending about 10 years with Pearson. It was during this time that I got a very close look at India's education ecosystem and realised where the biggest gap was.
My team, the co-founders, Saraswathy R, and Gowri Mahesh. They bring their own expertise - Gowri from training and technology, and Saraswathy from education. We are three co-founders, working together as a tight unit.
My personal motivation? It all came from my roots. I am from Trivendram, Kerala, and I had my entire education in a government school and then government college. I went to one of the top government engineering colleges in Kerala, but back then, the teachers spoke only in Malayalam. When I came out, I had a good background in English grammar and written skills, but my spoken ability? It was far, far away from what the job market required. I can completely empathise with the struggle. That’s why I can relate to what students in government schools go through today.
Give us the ‘Problem Overview’ you are trying to solve
Here’s the harsh reality of our country: We have about 15 lakh schools and a whopping 30 crore children in the K-12 sector. If you count higher education and skilling, we are looking at 60 crore learners!
We define the education market in India using a pyramid based on the fee structure. The top-end schools charging ₹1 lakh or more per year get everything. But the middle and the bottom of the pyramid (schools charging ₹25,000 up to ₹90,000, and of course, government schools) are left struggling.
The fundamental problem we are trying to solve is that the middle and bottom layers do not have access to the necessary resources. This directly impacts the quality of education, their learning outcomes, and eventually, their ability to earn a better living.
Specifically, we are solving the critical problem of English communication skills. I can't agree more, English is the new coding language in India. The biggest hurdle? We don't have enough English teachers in schools or colleges. This lack of skill is what often dictates whether a student gets paid ₹10,000 per month or ₹50,000 per month in their first job.
What motivated you to solve this problem?
My own background, graduating from a government college with poor spoken English, helped me realise the gravity of the issue. We saw that the quality of education in government schools has dropped drastically over the last 25 years. We need a strong foundation in schooling, but the infrastructure often isn't great.
We realised that the only way to solve this massive problem for the growth of learners is by bringing in technology. That’s why we left our comfortable jobs - to bring in technology and high-quality education together and make a real impact on these people's lives.
Give us a brief ‘Solution Overview’ - The AI Use Case
The solution is Tara AI. Very simply put, Tara is an autonomous voice technology teacher.
Think of how an Alexa speaker works: you speak, your voice converts to text, the program processes the text, and the result (text) is rendered back to you as voice. We rely on the Android app or Amazon Alexa to do this speech-to-text and text-to-speech conversion.
But beyond that, we have brought in eight levels of English communication programs. We have coded the grammar, correction logic, and feedback mechanism into the software, making Tara an autonomous voice teacher.
Here’s the twist: the learner is not asking questions to Tara. It's the other way around. Tara teaches, asks questions, and the learner has to respond.
Why did we use AI? Because finding enough human teachers to teach English to 30 crore children is impossible. If it were possible, the current education system would have already solved it!. AI allows us to give a good teacher assistant to the existing teacher or replace the lack of an expert teacher entirely. This autonomous voice technology is perfectly suited for teaching languages and non-mathematical subjects.
Tell us about - How AI is Actually Utilised?
We utilise what’s called Narrow AI - similar to a 'Hey Siri' or Google Assistant function.
The AI use case breaks down into two main parts:
- Platform Base: We built a skill called 'Tara' on the Amazon Alexa platform. We also created a dedicated app on the Google Play Store using standard Android technology. This is where we rely on the system (Alexa or Android) for the basic speech-to-text and text-to-speech services.
- Custom Logic: Beyond the basic voice conversion, we have built the specialized educational logic. This includes the Natural Language Processing (NLP) specifically trained to handle Indian accents, and the entire English grammar part, which is coded into the software. This ensures the correct feedback is provided - for example, explaining why "I like banana" is wrong and "I like bananas" is correct.
What ‘Challenges & Failures’ you faced in this project?
The main challenge was always scalability. We started before COVID, trying remote teacher training, but quickly realised that the number of expert English teachers available was too low to scale the impact.
The failure to scale using traditional methods pushed us toward tech. Our breakthrough came around 2019-2020 when two major things happened: internet prices dropped drastically (thanks to the dongle revolution) and the Alexa Echo Dot became popular. We put those two ideas together.
Our Version 1 program was basic. We did pilot runs in Tamil Nadu and learned a lot from the ground about how to make it effective in a real classroom. Those learnings went into Version 2 and then Version 3, expanding the scope and delivery methods. Now, we are working on an MoU with the Ministry of Education to bring NCERT textbooks into Tara, which shows how far we have come.
AI Readiness Skills
Name the persons or organisations you collaborated with to make this project successful.
Collaboration is absolutely vital. I mean, we are talking about impacting 30 crore students; we cannot do this without government help.
We have collaborated extensively:
- Ministry of Education, Delhi & NCERT: They have been extremely supportive because Tara aligns with their National Education Policy (NEP) goals, such as using the mother tongue for learning, and they appreciate that Tara is a no-screen solution, which helps with digital detoxification.
- RDPR (Rural Development and Panchayat Department), Karnataka: They are helping us roll out the program in libraries.
- DPWD (Department for Empowering People with Disabilities) and NAPD (National Institute for People with Visual Disabilities), Dehradun: Because our solution is fully auditory, it is a perfect fit for visually disabled people, and these ministries are supporting us.
- We also rely on CSR funds from various organisations, which helped launch our pilot in Maharashtra.
Honestly, I’m a huge, huge fan of LinkedIn. A lot of our connections and collaborations, including accidental ones like the discovery of the use case for visually challenged people, happened because of these networks. Without collaboration, we wouldn’t be where we are today.
Do you think your solution is the best one for the stated problem? If yes, then why? If not, what is the best solution?
If you look at the fundamental problem - teaching somebody a language skill, a non-mathematical skill - Tara is a perfect fit. It successfully provides a structured learning track across eight levels, teaching pronunciation, phrases, and sentences.
But is it the best solution? Not yet, because there is always space for improvement.
Our current design is excellent for structured teaching, but it struggles with open-ended or non-standard questions. For instance, if a student asks: "Why should I say 'an orange jug' and not 'a orange jug'?" - our system may not handle that open-ended follow-up question perfectly.
To solve the next level of skilling, like preparing someone for a complex job interview, we need more power. The best solution, in the near future, would be to integrate an LLM (Large Language Model) or experiment with a Small Language Model (SLM) to handle those nuanced, open-ended learner questions.
What self-learning did your team do to understand the problem better, or design the solution, or for any other aspect of the project? Also, please specify the sources (if possible) from which you learnt?
Like everyone else, we used the modern AI tools - a lot of ChatGPT and Gemini. Especially for pulling statistical data and asking follow-up questions to understand the market better.
But here’s the key insight: while online learning helps, the larger and more critical learning happens offline. You have to be on the ground, talking to teachers and students, to truly understand the real, messy problems they face. That on-ground journey is invaluable.
What efforts did you make to popularise your solution on the Internet?
We definitely use the internet, primarily LinkedIn, to reach out to ministries, ministers, MLAs, and MPs. While we are a small company and can’t spend millions on marketing, we focus on maintaining our reputation.
Reputation isn't just about follower count; it's about being positioned as the expert in your domain. We use our online presence to demonstrate our expertise and align ourselves with key stakeholders in the education space.
Impact so far & vision for the future
The impact so far is amazing. We have touched about 70,000 students across India. We have successfully delivered about 3.5 million hours of teaching on the Tara platform.
For the future, the vision is huge. The pilot launch in Sangli district, Maharashtra, is already planned to increase from 3 schools to 100. This will bring in tens of thousands of students quickly.
By April or May, we expect to launch the NCERT-based program, which will be a massive collaborative effort with the Ministry of Education. We are targeting lakhs, and hopefully, a few million students across India next year. Plus, we are expanding our language base from English to include Hindi, Tamil, and possibly Odia, aligning with the NEP vision.
Your advice to fellow innovators
My first piece of advice is: I don't think anyone is truly capable of giving advice because our experiences are all different.
But based on what I see in the industry, here is a warning: Adding an interface to ChatGPT or Gemini is NOT an AI product. That's like making a front-end screen for Google Search; you are not solving anything genuinely new.
If you are an innovator, your focus must be on building something unique that adds genuine value. We didn't invent speech-to-text, but we solved a painful problem - the lack of expert teachers - by turning existing AI technology into an efficient, active teacher assistant.
Focus on applications of AI that solve real problems, like creating a ChatGPT equivalent for Indian languages, or using existing tech to innovate where a solution is missing.
Mentor, Support & Inspiration
I can’t name one single mentor. I am thankful to hundreds and thousands of people - colleagues, friends, even strangers on LinkedIn, who constantly exchange feedback and guide us (like one person who messaged me about a huge grant opportunity today!).
But the true, core inspiration? It is the biggest kick we get when we see a learner in a Tier 2, Tier 3 town, or a village school confidently say some sentences or words after using Tara. That confident voice - that is the inspiration. That is why we are aligned with this purpose.
Kudos to the whole team of Learning Matters with Tara AI, and for proving India’s Got Intelligence. Let us together make India AI Ready.
