Rime Secures $24M Series A to Enhance Enterprise Customer Support

Voice AI startups have primarily unlocked value by managing enterprise calls across sales, marketing, and customer support. Large organizations are delegating call handling to voice model developers such as ElevenLabs and Deepgram, infrastructure providers like Vapi, Retell, and LiveKit, and specialized customer support firms including Decagon and Sierra.

San Francisco-based Rime aims to differentiate itself in this competitive landscape by leveraging voice AI models trained on proprietary conversational data it records, thereby minimizing the customization burden for its clients.

Established in 2022 by former Stanford PhD student Lily Clifford, ex-Amazon Alexa engineer Brooke Larson, and Stanford engineer Ares Geovanos, Rime has established a recording studio in San Francisco. This facility allows the company to gather its own conversational data instead of relying on web-scraped audio.

The startup specializes in tuning its voice models to accurately pronounce brand names and industry-specific terminology. By employing a phoneme-based architecture that adapts to various pronunciations, Rime ensures customers do not need to retrain models for their specific sectors.

On Wednesday, Rime announced a $24 million Series A funding round led by M13 Ventures. Existing investors, including Twilio Ventures, Corazon Capital, and Unusual Ventures, also participated in the round.

Clifford noted that despite advancements in voice AI, enterprises continue to favor legacy IVR systems because current AI voice technology has not yet matched the effectiveness of traditional IVR implementations.

“Voice technology has not yet reached the point where it can automate most enterprise phone calls. While LLMs have simplified the creation of functional voice applications, they have not transformed the user experience. Interacting with a voice AI agent remains less than compelling for end users; it is essentially a new IVR with improved audio quality,” she explained.

Initially, the startup utilized a pipeline of distinct models for speech-to-text, text-to-speech, and large language processing. However, it is now pivoting toward advanced speech-to-speech models to minimize latency, enhance turn-taking dynamics, and mitigate issues such as background noise. This streamlined approach also reduces the need for complex orchestration, allowing the company to avoid managing multiple disparate models.

Rime serves clients in the food service, healthcare, airline, and fintech sectors. The company asserts that its specialized training data and model optimization lead to longer customer call durations, which has helped secure enterprise contracts with major clients like Mayo Clinic, Dialpad, Upstart, and Asurion.

With this new capital, Rime plans to expand its workforce of 35 employees, focusing on hiring for model development, engineering, and partnership roles. The company recently appointed Rafael Valle, formerly of Meta Superintelligence Labs and NVIDIA’s applied deep learning audio research team, as its Chief Scientist.

“Companies like ElevenLabs are evolving into orchestration and application layers, competing directly with firms like Sierra and Decagon. There is significant technical work remaining, and Rime’s focus on delivering low-latency, high-reliability models within regulated environments distinguishes it,” Morgan Blumberg of M13 told TechCrunch.

The company previously raised $5.5 million in a seed round last May. As part of this latest funding, Blumberg is joining Rime’s board of directors.