Smallest.ai Secures $13M Series A to Build Real-Time Human-Like Voice AI

Voice-AI startup Smallest.ai raised $13 million to develop a compact model that can converse without perceptible pauses, aiming to make AI sound indistinguishable from a person.

Smallest.ai, a startup launched in late 2024, announced a $13 million Series A round to fund its effort to create a small voice model that processes, thinks, and speaks simultaneously, eliminating the awkward pauses typical of current AI agents. Founder and CEO Sudarshan Kamath explained that the model operates like a human interlocutor, interrupting when necessary and handing off complex queries to a larger foundational model for brief research.

The round was led by Seligman Ventures, with participation from Sierra Ventures and 3one4 Capital, pushing total capital raised to more than $21 million. The company’s solution provides a real-time intelligence layer for focused conversations, supports multiple languages and accents, and performs in noisy settings. Customers such as RingCentral and Truecaller already use the technology, and Smallest.ai positions itself against competitors like ElevenLabs, Cartesia and regional player Sarvam, emphasizing enterprise-focused, real-time voice agents.