With low data costs and increasing appetite of AI tools, the country has become central to the global race for AI user acquisition. It was last year when artificial intelligence (AI) giants including OpenAI, Google and Perplexity began offering free access to premium AI services in India, targeting a market of nearly one billion internet users.
With low data costs and increasing appetite of AI tools, the country has become central to the global race for AI user acquisition. Free plans are being used to build habits, collect data and build engagement in the new scale-driven market.
While much of the debate around India’s AI journey has focused on whether the country can build its own foundational models as compared to the West, that narrative is now taking a different route.
Several homegrown startups are now catering to the massive numbers of price-conscious users profitably, by deploying automated solutions across price-sensitive segments in education, skilling and speech therapy. Ultimately, tapping into new avenues for monetisation.
Reducing per-user costs
The demand in services such as spoken English training, academic tutoring, skilling and speech therapy have long seen strong interest across the country, particularly beyond metro cities.
What restricted scale historically was the linear nature of service delivery, which means more users required more teachers or specialists, pushing up costs.
Now, AI is breaking that equation. Companies are using it to automate the labor-intensive parts of service delivery and serve bulk users without their costs climbing at the same pace.
Spoken-English learning is one of the clearest examples.
One such startup building along the same lines is Kae Capital-backed Supernova, an AI-led edtech platform which helps users to build stronger spoken English and communication skills.
“The demand for spoken English has always been there. What’s been missing is a practical, affordable way to serve it,” Maharishi R.B., co-founder & CEO, Supernova told ETEntrepreneur.
Its core offerings include a 24/7 AI tutor that delivers interactive and personalised feedback on pronunciation, grammar and vocabulary. A human tutor for 20 minutes daily costs ₹5,000-10,000 per month, which is out of reach for most users in India.
“You only get better at speaking by actually speaking–that requires one-on-one practice. With AI, we’ve changed that equation. Same one-on-one practice, at ₹200-300 per month. Our cost to serve is less than ₹1 per day per learner,” Maharishi added.
The company has scaled to over $11 million in ARR (annual recurring revenue) by helping users learn English. Some of its other solutions include simulated conversations for users to enhance their speaking skills, along with pronunciation feedback and practice sessions.
Another such startup is Gurugram-based SpeakX.ai, which is building its offerings along similar lines.
“In India, many people want to learn spoken English. But most companies could not earn well because they needed human teachers. More learners meant more teacher hours, so costs kept going up. With AI, this changes. It can teach and practice with many users at the same time,” said Arpit Mittal, Founder & CEO at SpeakX.ai.
The firm raised $16 million in a pre-series B round led by WestBridge Capital, with participation from Elevation Capital and Goodwater Capital in October last year.
“This lets us serve large demand, keep prices affordable for Tier-2/3 users, and still keep good margins because we don’t need to keep hiring teachers to grow,” Mittal added.
When AI doesn’t miss pronunciation errors
Traditional language and skills training models rely heavily on live tutors, where quality varies by instructor, time of day and learner volume. One of the criticisms of large-scale education models has been inconsistent quality, which can be enhanced with AI, added the founders.
Tutoring depends heavily on individual instructors, where outcomes can vary by teacher or the learner volume.
“A human tutor might miss two out of five mistakes in a conversation. AI catches all five, every time, and gives precise feedback,” said Maharshi.
The firm has developed a proprietary five-step learning framework with defined rubrics, which refers to the scoring tool used to evaluate student work, to assess student’s progress.
SpeakX.ai follows a similar business model of automation along with human professionals in curriculum design and quality review.
In non-metro regions, the platform’s monthly plan is priced at around ₹199, positioned as roughly ₹6 per day.
“We keep our team deliberately small and use human effort only where it makes the biggest difference… AI manages everyday learning… Humans step in only for important tasks like updating the curriculum, reviewing quality, and handling rare cases,” added Mittal.
What investors are saying
Similar AI-led business models are being led by startups who are looking beyond language learning skills.
One such example is Iyaso, which offers an app called Eloquent, that provides AI-based, self-guided speech coaching for people who stutter. Similarly, Doubtnut uses image recognition and automated explanations to answer mathematics and science questions.
Similarly, Chennai-based Disprz, which operates an AI-led learning and skilling platform, currently serves over 1.7 million users globally.
Several investors believe the larger opportunity lies in removing human capacity as the primary constraint.
“The real value of these AI-led startups lies in their ability to monetise large, price-sensitive user demand by removing human capacity as the limiting factor. In India, demand has always been abundant, but traditional service models could not scale affordably. AI enables automation, continuous learning from data, and standardised decision-making at near-zero marginal cost,” said Somshubro Pal Choudhary, Co-founder & Partner at Bharat Innovation Fund.
He added how much like UPI and digital KYC standardised access at population scale, AI is now standardising intelligence and execution across sectors, making large markets economically viable for startups.
“Traditional English courses with a skilled teacher can cost five to ten thousand rupees and offer just a few hours of interaction each week. By contrast, an AI-powered English teacher is accessible anytime on a user’s phone for a fraction of the cost,” said Kushal Bhagia, Founder & Partner at All In Capital, who has invested in Supernova.
In India, digital businesses across e-commerce and mobility have long operated on low average revenue per user while serving large volumes.
AI startups are now entering a similar environment, where affordability and cost per interaction matter as much as model performance.
“India’s AI winners won’t be the ones with the best models. They’ll be the ones who figure out how to profitably serve a billion users at ₹10 per interaction. That’s a cost engineering problem as much as an AI problem,” said Tushar Jain, angel investor and Founder of IronLabs AI.
As AI deployment costs continue to fall, these startups are placing a calculated bet on whether high-volume, low-price models can generate profits. And if they pull it off, it could reset the playbook for AI businesses in India.
Leave a Reply