While artificial intelligence agents are increasingly adept at resolving customer support inquiries, the distinction between human and machine interaction remains readily apparent to most users.
Smallest.ai, a startup established in late 2024, posits that the next significant advancement in voice agent technology will not stem from accelerating large language models, but rather from employing more compact, specialized models engineered for authentic human conversation. In essence, the company aims to render interactions with an AI agent indistinguishable from speaking with a person.
To achieve this, Smallest.ai is developing a concise voice model designed to emulate the human process of information handling: listening, thinking, and speaking concurrently.
"While I’m speaking to you, you’re already thinking, and you might interrupt me if I talk for too long," Sudarshan Kamath (pictured left), founder and CEO of Smallest.ai, explained to TechCrunch, emphasizing that this dynamic is precisely how the startup's model is engineered to function.
To propel this ambitious mission, Smallest.ai has successfully secured $13 million in a Series A funding round. This round was spearheaded by Seligman Ventures, with additional participation from Sierra Ventures and 3one4 Capital. This latest capital infusion elevates the startup's total funding to more than $21 million.
Kamath elaborated on the limitations of current AI: "The way an LLM works is you give it an entire prompt, and then it starts thinking." He noted that while such latency is acceptable in text-based chats, even a brief pause in a voice conversation feels unnatural. "If you think about how we are talking, I’m not giving you like a large clipping of my audio, and then you start thinking."
Kamath envisions a future where all AI agents will operate using a dual-model architecture: a compact voice model dedicated to real-time interaction, complemented by an "offline" large language model that can be invoked as needed to tackle more complex challenges.
In contrast to expansive foundational models, Smallest.ai maintains a strict focus on voice-specific intricacies. This includes adeptly managing diverse accents, supporting dozens of languages, and performing effectively in noisy environments.
The startup's current clientele includes prominent companies within the voice technology sector, such as RingCentral and Truecaller. Kamath indicated that any customer support entity, including emerging players like Sierra and Decagon, represents a potential customer for Smallest.ai.
When questioned about why a well-resourced AI customer support company might not develop its own voice model, Kamath asserted that for such startups, becoming "extremely good at doing voice is a distraction from their core business."
Smallest.ai operates in a competitive landscape, contending with established voice AI leader ElevenLabs, as well as Cartesia and regional competitors like Sarvam, which specialize in local languages.
While some competitors apply voice AI to broader applications such as audio dubbing and podcasting, Smallest.ai concentrates exclusively on real-time conversational voice agents for its enterprise customers.
"We want our models to break the Turing test," Kamath declared. "You should speak to our model and not know it’s AI or human. That’s the sole focus of the company."
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.