Sometime around 2016, Mayank Goyal and his co-founders at ChatterOn noticed something odd. Their DIY chatbot platform - built to let anyone spin up a conversational bot - was getting used in a way they hadn't planned. Coca-Cola was using it. Vodafone was using it. Not for marketing, not for customer service. They were building internal HR bots. Quietly. On their own time. Because they had a problem and nobody was solving it properly.
That observation changed everything. The founders - three IIT Delhi batchmates who had been building together since 2011 - realized they were sitting next to the real opportunity, not on top of it. In August 2017, they shut down the general chatbot play and went narrow: enterprise employees talking to AI for support. They called it Leena AI.
"Our platform is like Siri for employees - answering their questions and helping them get things done at work."- Mayank Goyal, Co-Founder & Chief Scientist, Leena AI
Mayank's specific domain within that vision is the hard part: the AI itself. As Chief Scientist and Head of AI, he is responsible for what happens inside the models, the architectures, the reasoning systems that make Leena AI actually useful inside complex enterprise environments. While his co-founder Adit Jain runs company operations as CEO and Anand Prajapati handles engineering as CTO, Mayank owns the intelligence layer.
That division of labor matters. Building enterprise AI is not the same as building consumer AI. An employee asking about their maternity leave policy needs a different kind of precision than a chatbot suggesting a restaurant. It needs to be correct, compliant, context-aware, and connected to 30 enterprise systems at once. Building that - reliably, across 40+ languages, for Nestle and Airbnb and the Reserve Bank of India simultaneously - is the technical mountain Mayank has spent the better part of a decade climbing.
The climb had obvious milestones. Y Combinator's Summer 2018 batch. An $8M Series A from Greycroft in 2019. Then in September 2021, a $30M Series B led by Bessemer Venture Partners, joined by B Capital Group - Eduardo Saverin's fund - and Greycroft returning. By that point, Leena AI had proved it could serve enterprises at scale. The funding was not for survival; it was for acceleration.
What acceleration looks like: 500+ enterprise clients, 3+ million employees, 90+ countries, 40+ languages, and three consecutive years as a Gartner Emerging Market Quadrant Leader for AI Knowledge Management Apps. The Gartner recognition is the kind that makes procurement departments pay attention. In enterprise software, being on a Gartner quadrant is not decoration - it is a purchasing signal.
The product itself has evolved considerably from the early HR chatbot days. Leena AI now runs what it calls an agentic architecture - autonomous AI colleagues that don't just answer questions but complete multi-step workflows across IT, HR, Finance, and Operations. Gavin handles HR. Miles handles IT. Harper handles Finance. Each is a pre-built AI colleague that enterprises can deploy and customize, or they can build new ones entirely through the AI Colleague Studio launched in October 2025.
The voice layer arrived in July 2025 - employees can now speak to their AI colleagues rather than type. That shift matters less for novelty and more for accessibility: in global operations where English is a second language, or in plant floors where keyboards aren't practical, voice changes who can benefit from the system.
The numbers Leena AI reports are striking. Customers see 70%+ efficiency gains. Manual employee support interactions drop by 70%+ post-deployment. IT and HR teams project 50% productivity boosts. These are the metrics that move enterprise deals, and they suggest Mayank's AI architecture is doing something real, not just generating plausible-sounding answers.
There is a particular kind of credibility that comes from building something that works inside Coca-Cola and the Reserve Bank of India at the same time. Regulatory environments differ. IT infrastructure differs. Cultural expectations of what AI should and shouldn't say differ. Maintaining performance across all of it requires the kind of deep, unglamorous engineering work that Chief Scientists do rather than announce.
Mayank and the Leena AI team are now operating in a landscape that has become significantly more crowded since 2017. Every major software vendor has an AI copilot. Every hyperscaler has an enterprise AI offering. What Leena AI has that most newcomers don't: seven years of data on what enterprise employees actually ask, what workflows actually matter, and what makes AI support trusted rather than tolerated. That training data advantage compounds quietly.