For much of the last decade, India’s edtech story was written in the language of disruption. Companies chased students directly, customer acquisition became an industry in itself, and valuations rose on the promise that learning could be unbundled from the classroom. It was a persuasive proposition while schools were shut and capital was plentiful. But as classrooms reopened and investors became more demanding, the sector was forced to confront an older truth about education: trust is not built in a matter of months, and learning outcomes are not created by downloads alone.
That correction has been sharp. Indian edtech funding, after peaking at $4.1 billion in 2021, fell to $321 million in 2023, with only a modest rebound visible in 2024. Globally too, edtech venture capital in 2024 dropped to $2.4 billion, its lowest level in a decade. Despite this downturn, a few green shoots are visible.
In this chastened market, Extramarks looks different from many of its peers. Founded in 2007 by Atul Kulshrestha, the NCR-headquartered company did not build its primary model around bypassing schools. It embedded itself within them through smart classrooms, curriculum-aligned content, assessment frameworks and teacher support. Today, the company says it reaches over 10 million students and 300,000 teachers across more than 20,000 schools in India and international markets.
For Ritvik Kulshrestha, who took over as MD and CEO in 2019, this institutional base is the company’s biggest strategic asset. “We have always believed that schools are the fundamental anchor of the education system,” he says. “Technology should strengthen the school itself and help create what we think of as a truly connected classroom.”
That view gives the company an interesting position in the post-pandemic edtech landscape. While several players are now trying to retrofit offline or hybrid models onto businesses built for direct-to-consumer scale, Kulshrestha is attempting the reverse: modernising a school-first foundation with AI, data and connected workflows.
Digital adoption on the rise
The distinction matters because India’s school system is too large, uneven and socially embedded to be disrupted by apps alone. According to the Economic Survey 2025-26, the school education system serves 24.69 crore students across 14.71 lakh schools, supported by over 1.01 crore teachers. UDISE 2024-25 data also shows that computer access in schools rose from 57.2 per cent in 2023-24 to 64.7 per cent, while internet connectivity rose from 53.9 per cent to 63.5 per cent. Digital adoption is rising, and for any edtech company this expands the ways in which it can support students in their learning journey.
The winners, analysts increasingly suggest, will not be those who sell technology as an escape from the school system, but those who can work within its complexity. HolonIQ’s Education Intelligence Unit has framed the investment reset as a shift from pandemic-era expansion to a more measured market in which investors are seeking ‘fundamentals over fluff’.
Kulshrestha’s own journey explains his focus on systems rather than just spectacle. An electrical and electronics engineering graduate from UCLA, he began his career at Amazon on the Checkout Engineering Excellence team, where he worked on shaping high-scale consumer experiences. That engineering exposure continues to influence his product thinking, as he believes that education cannot be solved only at the interface level. A school product must fit into the teacher’s preparation time, the principal’s need for visibility, the parent’s need for reassurance and the student’s need for continuous learning support.
There is also a second-generation leadership story here, though not the usual one of rupture. Founder and Chairman Atul Kulshrestha built the company around institutional trust at a time when smart classrooms were still a new idea in India. Ritvik has not abandoned that foundation. Instead, he is adding intelligence to it. Since 2019, he has led its evolution from a classroom digitisation provider into a comprehensive, school-integrated learning platform that supports the academic journey from in-class learning to after-school reinforcement and competitive exam preparation.
Its integrated AI layer, Extra Intelligence or ExtrAI, is built on proprietary academic content and real classroom usage data accumulated over the years. Unlike a standalone assistant, it operates within a defined academic structure, aligned to curriculum, pedagogy and assessment frameworks.
Over the next 5 years, schools will move from being time-bound institutions to continuous learning systems
“This is not AI as a feature. It is AI as an integrated layer that enables personalisation at scale, connecting teaching and learning into one continuous system,” says Ritvik.
That is perhaps its sharpest positioning in the AI moment we are living through. While much of the market has rushed to describe AI as a student copilot, this company describes it as a ‘connected ecosystem’ where schools, teachers, students and parents operate on a single platform.
The timing is significant. In May 2026, India launched an ‘AI Literacy for Teachers’ programme through Bodhan AI, a Centre of Excellence in AI for Education incubated at the Indian Institute of Technology Madras, with the goal of training one million teachers by 2027. Education Minister Dharmendra Pradhan said the initiative was about empowering teachers to become “architects of a future-ready India”.
That policy direction aligns closely with what Extramarks is trying to argue commercially: AI in education will scale only if teachers accept it as useful, safe and curriculum-aware. At MES International School, Pattambi, Kerala, Principal Shrilekha points to a practical shift.
“Teachers,” she says, “use tabs after a 30-minute lesson to conduct short assessments and understand how much students have grasped before moving ahead.” The value, in her telling, lies not in digitising a quiz for its own sake, but in turning classroom attention and comprehension into something a teacher can immediately see.
The challenge, of course, is not merely to put AI in a classroom. It is to make it contextually reliable. Indian schooling is fragmented across boards, states, languages, fee segments and pedagogical habits. Generic AI tools may impress in demonstrations but fail in school workflows if they do not understand curriculum progression or assessment patterns.
Technology should strengthen the school itself and help create what we think of as a truly connected classroom
Ritvik argues that generic models remain probabilistic and unpredictable in a classroom setting, while schools need a controlled environment where outputs are accurate, age-appropriate and curriculum-aligned.
AI-assisted evaluation
That contextual layer becomes particularly important during assessments. Subjective answer evaluation remains one of the most time-intensive academic processes in Indian schools. Extra Intelligence enables AI-assisted evaluation of handwritten responses using structured, rubric-based scoring aligned to board marking schemes. It can also generate dynamic question variations at a concept level, making copying ineffective while preserving assessment fairness.
The impact is also visible in how students understand difficult concepts. Aryaveer Gupta, a Class 9 student at Hyderabad Public School, Begumpet, Telangana, describes it as a blend of education and visual learning.
He adds: “In subjects such as Physics, Chemistry and Biology, graphs, graphics, experiments and demonstrations help students understand complex ideas more clearly. It makes teachers’ work efficient too because students can grasp concepts by seeing them before moving to tests and quizzes that bring greater clarity.”
That is where the company’s ambition begins to move beyond smart classrooms. It wants to become an operating layer for schools, connecting principals, teachers, parents, assessments and after-school learning. The phrase may sound expansive, but it reflects a broader shift in global edtech. HolonIQ’s 2025 Global EdTech 1000 noted that K-12 still accounts for nearly 40 per cent of its cohort, driven by tutoring, content, school support and classroom infrastructure.
For the company, the opportunity is sizeable, but so are the challenges. Institutional sales cycles are long. School budgets are uneven. AI adoption will require training, trust and clarity around data. And in education, technology cannot hide behind engagement metrics indefinitely. It must eventually show whether students learn better and teachers work smarter.
Ritvik appears aware of that need. That temperament may matter in the next phase of edtech. The sector’s first wave rewarded speed. The next may reward patience, credibility and integration.
“Over the next 5 years, schools will move from being time-bound institutions to continuous learning systems,” says Ritvik. “The role of the principal will become more data-driven. Instead of relying on periodic reports, they will have continuous visibility into academic performance, classroom effectiveness and teacher engagement.”
In a market once obsessed with replacing the classroom, Extramarks is making a quieter case for reinforcing it from within. The argument is not as flashy as disruption, but it may prove more durable. Education, after all, is not a category where trust can be acquired cheaply. It has to be earned lesson by lesson, teacher by teacher, school by school.