Smallest.ai, a startup founded in late 2024, has secured $13 million in a Series A funding round led by Seligman Ventures, with participation from Sierra Ventures and 3one4 Capital. The company’s mission is to revolutionize voice AI by developing ultra-fast, human-sounding voice models that enable AI agents to interact naturally and indistinguishably from humans. This latest financing brings Smallest.ai’s total funding to over $21 million, underscoring strong investor confidence in their approach to voice-based AI.

The startup’s innovation lies in creating smaller, specialized voice models designed to operate in real time by mimicking the way humans listen, process, and speak concurrently. Unlike traditional large language models that require processing complete inputs before responding, Smallest.ai’s technology reduces response lag, enabling more fluid and conversational exchanges. When the model encounters unfamiliar queries outside its knowledge base, it seamlessly hands off to a larger language model to “research” answers, replicating the behavior of a human customer support agent.

Smallest.ai focuses specifically on voice AI nuances such as accommodating diverse accents, supporting multiple languages, and functioning well amid noisy environments. Its clients already include voice-centric companies like RingCentral and Truecaller, with potential applicability for many customer support firms who might find building their own voice AI distracting from their main business goals. The startup faces competition from established players like ElevenLabs but differentiates itself by concentrating solely on real-time conversational voice agents rather than other voice AI applications like dubbing or podcasting.

CEO Sudarshan Kamath envisions a future where AI agents combine a small real-time voice model with an offline large language model to balance immediacy and complex problem-solving. Smallest.ai’s ultimate goal is to create voice AI that passes the Turing test in phone calls, making it impossible for users to distinguish AI from human agents. This pursuit highlights a significant leap forward in customer support technology and the broader adoption of AI in natural voice interactions.