Smallest.ai, a startup founded in late 2024, has secured $13 million in Series A funding to advance a new generation of voice AI built for natural, real-time conversation. The round was led by Seligman Ventures, with participation from Sierra Ventures and 3one4 Capital, bringing the company's total funding to more than $21 million.
Rather than relying on larger language models alone, the company is developing a compact voice model designed to listen, process, and respond in a way that mirrors human dialogue. The goal is to reduce the pause that often makes AI sound mechanical and to create conversations that feel fluid and immediate.
A Voice-First Approach
According to founder and CEO Sudarshan Kamath, the model is built to work in real time, handling speech as it happens instead of waiting for a full prompt before generating a reply. When a request goes beyond its core expertise, the system can shift the task to a larger model for deeper reasoning, creating a hybrid workflow that blends speed with broader intelligence.
Smallest.ai says this approach is especially useful for customer support and enterprise voice applications, where natural timing, accent handling, multilingual support, and performance in noisy environments can shape the user experience. The company already works with voice-focused customers such as RingCentral and Truecaller.
Kamath argues that future AI agents may combine two layers: a fast voice model for live interaction and a larger offline model for more complex tasks. With that strategy, Smallest.ai is aiming to make voice systems feel less like software and more like seamless conversation partners.
If the company delivers on that vision, the next wave of voice AI could redefine how people interact with digital assistants across work, services, and everyday life.
