India’s AI Revolution Is Here
BharatGPT isn’t copying ChatGPT—it’s quietly powering India’s public services in 14+ languages.
“I’ve never called myself a businessman. I’m still a techie. Still a problem solver.”
Ankush Sabharwal didn’t start out trying to build an AI company.
He started with a simple belief: if humans can talk to each other, they should be able to talk to machines.
Today, he’s the founder and CEO of CoRover.ai, the company behind BharatGPT—India’s first multilingual conversational AI platform, deployed across IRCTC, LIC, NPCI, and more.
But before that? He was a software engineer, a BFSI tech operator, and a kid who once took a 38-hour train ride to join his first job at Wipro.
This episode is a deep dive into how a grounded, tech-first founder built one of India’s most practical and widely deployed AI companies—without noise and buzzwords.
🎯 What You’ll Learn
Why real founders obsess over customer problems, not pitch decks
How BharatGPT is being built for India’s complexity, not Silicon Valley’s PR cycle
What it takes to sell conversational AI to India’s largest enterprises
How not to get disillusioned by fundraising, flashy demos, or overnight unicorns
The difference between “building a company” and “building something useful”
🎙️ About This Episode
Ankush Sabharwal is the founder and CEO of CoRover.ai, the company behind BharatGPT. His platforms support mission-critical conversations across banking, insurance, travel, and public sector applications for 100M+ users.
Before founding CoRover, Ankush worked in leadership roles at Wipro, Misys, and Altisource. He’s now focused on building India’s conversational infrastructure—one machine at a time.
How Ankush is wired differently?
“We don’t celebrate when the money hits the account. We celebrate when a customer says, ‘You solved my problem.”
🧭 From Wipro to Voice AI
In 2004, Ankush took a 38-hour train from Ranchi to Bangalore to join his first job at Wipro.
He didn’t know then that he’d eventually build a company powering India’s railways, banks, and tourism boards with conversational AI.
But what he did know—even back then—was that user experience matters.
He realized this while working with banks: everyone used the same software. The only difference was how customers felt when they interacted with the brand.
That insight—interaction > features—became the foundation for CoRover.
The BharatGPT Thesis
Most AI companies are chasing models.
CoRover is chasing outcomes.
Their belief is simple: every machine should talk. Whether it’s a washing machine, ATM, or CNC lathe—if it can compute, it should converse.
Before LLMs took off, CoRover was already building assistants using entity extraction, rule-based NLP, and speech recognition—in multiple Indian languages.
Their early chatbot didn’t even respond with open text. It gave three options for the users to choose.
It wasn’t perfect. But it worked.
And that’s what Ankush cared about.
“If I can talk to you, why can’t I talk to my machine?”
💼 Selling to Giants (as a Tiny Team)
It took CoRover over 2 years to close their first major enterprise deal.
They were a team of 3. No brand. No funding. Just persistence.
They pitched, followed up, flew to Delhi, sat through retenders—and lost.
Yes, lost.
But they celebrated anyway. Because they knew that building trust was harder than building tech.
And once that trust was built, they got in. IRCTC, LIC, and NPCI—all followed.
“The project got re-tendered and we didn’t win it. But we went to Goa and celebrated. That’s how much we believed in what we were doing.”
Product First, Ego Last
Ankush doesn’t have a “founder mindset.” He has a builder’s mindset.
He doesn’t fake confidence. He doesn’t posture.
When he’s unsure, he tells his team. When he feels stuck, he talks to customers. When asked about his entrepreneurial journey, he shrugs.
“I don’t think of myself as a businessman. I still think of myself as a techie.”
That authenticity shows in how he talks, hires, and sells. He’s not chasing the spotlight. He’s chasing product-market fit.
🇮🇳 BharatGPT doesn’t believe in the AI Hype
BharatGPT isn’t trying to beat OpenAI.
It’s solving problems for a different country.
It is building:
Multilingual voice-based assistants
Low-bandwidth AI deployments
Domain-specific bots (railways, insurance, tourism)
Interfaces that work for India’s diversity, not despite it
It’s not a demo. It’s live. In Hindi, Kannada, Bengali, Tamil, and more.
“The real India doesn’t want to talk to a chatbot. They want answers. In their own language.”
🎥 Watch the Full Conversation
👥 Your Hosts
Pushkar Singh
Pushkar Singh is a seasoned entrepreneur and investor committed to supporting founders in building outstanding businesses. He also conducts fundraising workshops, equipping founders with the knowledge and tools to navigate the complex landscape of startup financing.
Sandeep Balaji
Sandeep Balaji is a growth and go-to-market (GTM) expert, operator, and investor with extensive experience in the startup ecosystem. He is passionate about fostering a supportive community for founders and facilitating honest discussions about the challenges and triumphs of building startups in India.
CoRover isn’t trying to go viral. It’s trying to be useful.
And in the age of hype-first, value-later AI startups, that may be the most radical thing.
Happy Building.