Rime, a San Francisco-based voice AI platform, raised $24 million in a Series A round led by M13 on July 15. Twilio Ventures, Corazon Capital, and existing backers including Unusual Ventures participated. The round brings Rime’s total funding to just under $30 million, and comes as the company is handling nearly 100 million enterprise phone calls per month.

The raise is the latest signal that voice AI for business is moving from experimental to infrastructure. The question is no longer whether enterprises will deploy voice AI agents — it’s which ones they will trust for the calls that actually matter.

What Rime Actually Does

Most voice AI platforms run into the same problem: they mispronounce brand names, struggle with medical or legal terminology, and sound fine until they don’t. That failure mode is particularly damaging in healthcare and financial services, where a wrong pronunciation either confuses a patient or erodes trust at a critical moment in the customer relationship.

Rime attacked this differently. Rather than training on scraped web audio, the company built a recording studio in San Francisco to collect conversational data directly. This lets them control the quality, diversity, and domain coverage of their training set in a way that pure web-scraping can’t match.

Their technical architecture is phoneme-based, which means the system reasons about how words are actually constructed rather than memorizing surface patterns. The practical result is that enterprises can deploy Rime’s models for their specific industry without needing to retrain from scratch for every term that matters in their domain.

Who’s Using It

Rime’s customer list reads like a deliberate showcase of high-stakes environments:

• Mayo Clinic — where pronunciation accuracy in patient-facing voice interactions carries real consequences
• Dialpad — a major enterprise communications platform integrating voice AI natively
• Upstart — lending platform where voice agents handle loan-related queries
• Asurion — one of the largest device protection companies globally, handling millions of support calls

Together these customers represent the range of enterprise voice AI workloads: clinical information, financial services, complex support flows. That spread suggests the platform is genuinely multi-sector, not a healthcare specialist or a call center tool.

New Leadership for the Research Side

Alongside the funding, Rime brought on Rafael Valle as Chief Science Officer. Valle previously led audio understanding research at Meta’s Superintelligence Lab and worked on applied deep learning for audio at Nvidia. That’s a research hire that signals Rime is building toward more fundamental capability improvements, not just packaging existing model infrastructure.

Morgan Blumberg, Partner at M13, joins the board.

What the $24M Will Do

The company plans to expand its 35-person team with hires across model development, engineering, and partnerships. The investment will also go toward expanding their proprietary conversational dataset — the core competitive asset that underpins their accuracy advantage over generic voice AI providers.

What This Means for Business

The Rime raise lands at an interesting moment. Enterprise voice AI is consolidating around a short list of platforms that can clear a real bar: HIPAA compliance, industry-specific accuracy, and reliability at scale. The generic text-to-speech solutions that businesses bolted onto their workflows three years ago aren’t passing that bar.

Two dynamics are driving this:

First, the volume of AI-handled calls is growing fast. When a platform is handling 100 million calls per month and that number keeps growing, even small accuracy improvements compound into massive differences in customer outcomes and brand risk.

Second, voice AI is expanding beyond inbound support into proactive outreach, appointment scheduling, and follow-up flows — places where the stakes of sounding wrong or saying something incorrect are higher. That’s pushing enterprises toward platforms with stronger accuracy guarantees.

For businesses evaluating voice AI agents right now, the key question to ask any provider is where their training data actually comes from. Web-scraped audio has different quality characteristics than purpose-recorded conversational data, and that difference shows up in exactly the contexts that matter most.

The investment from Twilio Ventures is also worth reading as a signal. Twilio is deeply embedded in enterprise communication infrastructure. Their involvement suggests Rime is being evaluated as a potential component of larger enterprise telephony stacks, not just a standalone voice AI product.

---

Enterprise DNA builds voice AI employees for businesses through Omni Voice — from inbound knowledge discovery to admin automation and team communication. If you’re evaluating enterprise voice AI for your operations, book a discovery session to work through the options.