Copyright yourstory
In India, voice communication still dominates how businesses interact with customers, whether it’s banks following up on loans, recruiters screening candidates, or ecommerce platforms confirming deliveries. However, running large call centres in a country with 22 scheduled languages and numerous dialects is costly and inefficient. Traditional voice automation tools fall short in understanding local accents, switching between languages, or maintaining natural, human-like conversations. This is the gap Bolna AI set out to solve. Founded in 2024, the Bengaluru-based startup is developing a voice AI infrastructure that enables enterprises to automate conversations seamlessly, encompassing speech-to-text, natural language processing, and text-to-speech within a single integrated platform. Co-founder Maitreya Wagh, an IIT Delhi graduate, started his career at Bain & Company before joining opinion trading startup Probo in a founder’s office role, where he saw how fast-growing companies struggled to scale customer communication. Along with Prateek Sachan, a seasoned engineer who has worked with Zomato, BrowserStack, and Atlassian, he founded Bolna AI to make high-quality, affordable voice automation accessible to businesses of all sizes. “I saw how chatbots were taking off, but in India, people still prefer talking,” says Wagh, “As AI models got better and cheaper, I knew the same revolution would come for voice.” Building India’s voice AI layer Bolna AI serves as an orchestration layer, simplifying the complex backend of voice automation by managing call routing, latency, accent recognition, and real-time context memory. The Y Combinator-backed uses speech-to-text, LLM, and text-to-speech models to enable natural, human-like conversations. At the same time, its system intelligently selects and optimises the best AI models for each call. It uses both its own proprietary AI model and external platforms like OpenAI and Sarvam AI to ensure fast, accurate, and lifelike responses. The platform offers both no-code tools and APIs, allowing enterprises and developers to create and deploy voice agents for use cases like recruitment, customer support, and logistics. What sets Bolna apart is its per-minute pricing model based on voice agents’ usage. Businesses are charged only for usage, starting at Rs 7 per minute and dropping to Rs 3 for higher volumes. The system automatically optimises costs every six months by analysing usage data and improving efficiency through smarter routing and token usage. A growing customer base Operating on a B2B model, Bolna already has around 800 paying users, including GoKwik, Footwork, and Awign. Fourteen large enterprises currently use the platform for high-volume voice automation, while smaller businesses and independent developers use it for specific campaigns and pilots. Globally, Bolna’s closest competitor is VAPI, another YC-backed firm offering similar developer-focused voice APIs. In India, it competes with players like Sarvam AI, but Bolna differentiates itself by being model-agnostic, choosing the best speech model for each use case instead of locking clients into one provider. According to IMARC, the voice recognition sub-market stood at $462.8 million in 2024 and will grow to $2.98 billion by 2033 at a 23% CAGR. According to Wagh, India sees over a billion calls a day, 30% of which are business-related, “but we’re currently processing around one million calls a month, a fraction of the market but one with immense room for growth.” Looking ahead Bolna AI plans to grow its business with more Indian enterprises while also expanding into global markets like the US, Brazil, and Southeast Asia. The team of 8 people is developing its own AI models that better understand Indian languages, including regional names and numbers across different dialects. Also, as AI evolves quickly, the startup aims to become the main platform for voice automation, helping businesses use lifelike voice agents in multiple languages at a large scale and much lower cost. Bolna AI is part of YourStory’s Tech30 cohort—a selection of India’s most promising startups of 2025—unveiled at TechSparks Bengaluru.
