India’s edtech sector is entering its second act, one marked by an aggressive shift offline. After years of digital disruption, leading players are investing in physical centres to rebuild trust, improve outcomes, and achieve sustainable unit economics.
PhysicsWallah (PW), once a purely digital success story, is preparing a ₹3,480 crore IPO to fund offline expansion and acquisitions such as Utkarsh Classes, strengthening its reach in tier-II and tier-III cities. Meanwhile, UpGrad and Unacademy are reportedly in talks for a $300–400 million merger — a move that signals large-scale consolidation as both seek deeper offline footprints.
UpGrad has been expanding into physical learning centres and enterprise programs, while Unacademy, after restructuring for profitability, expects 70% of its centres to turn profitable in 2025.
Even Vedantu and SpeakX are following suit, using recent funding rounds from Accel and WestBridge Capital to grow hybrid and classroom-based models.
The message is clear: despite their digital DNA and efforts towards AI, India’s edtech leaders are betting big on brick-and-mortar. AI now plays a supporting role — bridging digital precision with classroom credibility, as the sector learns that the future of learning still has a physical address.
AI and Offline Evolution Through the Pandemic
The pivot to offline and hybrid learning began soon after the pandemic. During 2020–2021, India’s edtech sector boomed as platforms like Byju’s, Unacademy, Vedantu, and PhysicsWallah scaled rapidly with AI-driven systems and virtual classrooms. But, as schools reopened in late 2021, engagement fell. Parents sought credibility, students wanted peer interaction, and teachers struggled to sustain attention online. The shift forced edtech firms to transition from their growth-at-all-costs model to sustainable, outcome-focused models.
Even a physical AI-based tutor won’t be able to replace a classroom, because offline teaching provides students emotional and mental support, they need to interact with the teacher, said Suraj Biswas, founder and CEO of Assessli.
Assessli is building, it claims, “the world’s only LBM (Large Behavioral Model) — an AI foundation model trained on genomic, neuropsychological, and behavioural data.”
Biswas added that edtech firms are routing to offline models because they need to sustain themselves and keep the revenue rolling.
By 2022–2023, the offline pivot gained pace. Byju’s launched Tuition Centres, Unacademy opened hubs in Kota and Bengaluru, Vedantu rolled out hybrid learning pods, and PW expanded into tier-II and tier-III cities through Utkarsh Classes.
AI shifted from automation to classroom support. Unacademy’s Airlearn offers AI-driven language practice in English and regional languages, PW uses AI dashboards to personalise guidance, and Vedantu’s WAVE 2.0 monitors engagement to adapt teaching in real time. Byju’s added computer-vision tools to track participation. These examples show how edtechs are refitting their AI for the classroom.
This model can work because offline teaching and AI are complementing each other to shape the learning experience, said Shantanu Rooj, founder and CEO at TeamLease Edtech. “AI gives scale, personalisation, and 24×7 academic support. While offline centres offer trust, community, and structure, especially for high-stakes learning,” he added.
Rooj said that recent industry studies show that hybrid models improve completion and engagement versus purely online or purely physical formats. Over 70–80% of institutions are experimenting with some form of blended delivery.
“For a company like ours that is deeply committed to online and work-linked learning, offline touchpoints make sense as they strengthen outcomes, not replace digital-first efficiency,” he added.
The Profitability Game
Edtech companies are moving towards offline learning as many students still prefer face-to-face classes, especially for competitive exams and certification courses. Even though fewer people take these offline courses compared to digital ones, companies focus on them because students stay longer (lifetime value of customer) and the cost to get each customer is lower (customer acquisition cost).
Biswas elaborated to explain: These firms struggle for profitability as online the CAC is very high compared to offline. A lot of marketing and retention goes into online as the CAC to LTV ratio is small, which should ideally be 1:7—if CAC is 10%, LTV should be around 70%.
“Otherwise, if I spend ₹3,000 or even ₹500 to get one user and only earn ₹3,000–₹5,000, the ratio isn’t that good. That’s why offline is always better. Of course, it depends — user to user, organisation to organisation, course to course — but this is the ideal case.”
He added that retention is much lower in online learning because students have many options. “In online, the cost is around ₹4,000–₹5,000, so students don’t mind subscribing to multiple platforms — Allen, Unacademy, Vedantu — or even using free YouTube classes,” he said. Yet, having digital operations helps in building brand presence in tier 2 and tier 3 cities where students lack infrastructure.
AI and the Blended Future
Acquisitions of AI-led edtech platforms by offline centres, or vice versa, are only one way forward, said Rooj. He said that offline players can bring AI into their systems through partnerships, in-house solutions, or platform integrations that support adaptive practice, doubt-solving, vernacular learning, and performance analytics.
According to him, the real advantage will come from combining three core elements: human context through mentors and counselling, AI infrastructure that enables personalisation and real-time feedback, and visible outcomes linked to grades, exams, or jobs. Centres that act as high-trust spaces powered by intelligent digital systems, he said, will hold a stronger edge than those treating AI as a superficial add-on.
Rooj claimed that traditional “content plus testing” models are losing relevance as exam preparation shifts toward conceptual mastery, analytics, and clear career pathways.
Biswas added to this by dividing learners into two groups—offline and online—each using AI differently. Offline learners rely on mentoring and structure, but increasingly benefit from digital augmentation. In classrooms, AI is automating question papers, evaluations, and performance tracking.
Companies like Extramarks are building B2B SaaS tools to help teachers optimise classes and analyse learning outcomes. He noted that while teachers use AI to create questions and students use it to draft answers, there is now a need for balanced tools that assist both sides without bias.
Looking ahead, Biswas said he expects “physical AI companions” to enter classrooms—interactive tutors capable of explaining concepts, drawing visuals, and generating content dynamically across formats such as text, diagrams, and videos.
For online learners, the biggest challenge is retention in an environment overloaded with content. While AI enables personalised learning and content generation, gamified experiences are emerging as the real differentiator because they keep a fickle and easily distracted audience engaged and motivated to stay longer on a platform. These operations might use AI and increase operational costs.
Biswas believes the next major opportunity lies in hybrid learning models across tier-II and tier-III cities, where demand is high but access to quality education remains low. Institutions that blend offline trust with AI-driven personalisation, he said, are best placed to lead this next phase of growth.
The post Edtech’s Offline Shift reshapes AI Usage appeared first on Analytics India Magazine.
