Guava Raises $12.5M Seed Round Led by Acrew Capital for Enterprise Voice AI

Guava has raised $12.5 million in seed financing to scale its developer platform for AI-powered voice agents, as enterprises push automated customer conversations into production.

Acrew Capital led the round, with participation from Flourish Ventures, WillowTree Ventures, Clocktower Ventures and existing investor Commerce Ventures, the Los Angeles-based company announced.

The funding will bankroll continued development of Guava's proprietary voice AI models and the creation of a go-to-market team aimed at enterprises and business process outsourcing providers.

What does Guava actually build?

Guava launched its first public model, Daytona, in September 2026. The company says more than a dozen enterprise customers already use the technology to handle tens of thousands of calls daily.

The platform builds on research from Gridspace, a voice AI laboratory whose models were trained on more than 10 billion minutes of conversations from regulated call centers.

Unlike most voice agent platforms, which stitch together speech recognition, language processing and speech synthesis from separate providers, Guava develops and operates its own models across the entire voice interaction. The company says that approach improves response times, reliability and consistency, while giving developers tighter control over how automated agents handle customer conversations.

Guava supplies software development tools for Python and TypeScript, letting businesses embed voice agents into existing applications and workflows using code-based business rules rather than leaning primarily on prompts.

Why did investors back it?

Co-founder and CEO Omri Dahan says the pitch rests on a consistent complaint from buyers. "Nearly every company we spoke with was dissatisfied with the voice AI they'd tried. It sounds fine in a demo but breaks down on real, high-stakes calls," Dahan said.

Dahan previously served as chief revenue officer at payments technology company Marqeta. He co-founded Guava with Anthony Scodary and Nico Benitez, technical leaders from Gridspace.

How fast can customers deploy?

Guava says developers can build a working agent in under five minutes using its free offering. Enterprise customers typically deploy their first live regulated call within 14 days, according to the company.

The new capital will also fund advances in three technical areas:

• On-premises deployments
• Edge devices
• Full-duplex conversations, which allow AI agents to listen and speak simultaneously

What's the market context?

The investment lands as businesses ramp up evaluations of voice AI for customer service, appointment scheduling and other high-volume interactions. That shift is creating demand for systems that can handle complex conversations reliably at scale — the specific gap Guava says Daytona fills.

With $12.5 million in hand and a go-to-market team in the works, Guava's near-term test will be converting its 14-day deployment speed and end-to-end model stack into enterprise accounts beyond its current dozen-plus customers.