Guava, a developer platform for enterprise voice AI agents, has raised $12.5 mn in seed financing led by Acrew Capital to accelerate product development and expand adoption among businesses and business process outsourcing (BPO) providers.
Flourish Ventures, WillowTree Ventures and Clocktower Ventures participated in the round, alongside existing pre-seed investor Commerce Ventures.
- Guava secured $12.5 mn in seed funding led by Acrew Capital, with participation from Flourish Ventures, WillowTree Ventures, Clocktower Ventures and Commerce Ventures.
- Its Daytona voice AI model serves more than a dozen enterprise customers and processes tens of thousands of calls daily.
- The funding will support AI model development, enterprise sales expansion and deployment capabilities, including on-premises systems and simultaneous speech processing.
The Los Angeles-based company plans to invest in its proprietary voice models, developer tools and a new go-to-market team focused on enterprises handling large volumes of customer interactions.
Guava’s first publicly available model, Daytona, launched in September 2026 and is already used by more than a dozen enterprise customers, handling tens of thousands of calls each day.
Guava targets enterprise demand for reliable voice AI agents
Guava is developing voice agents designed for customer interactions in industries where accuracy, reliability and operational controls are particularly important.
The platform builds on more than a decade of voice AI research conducted by Gridspace, which trained its models using more than 10 bn minutes of live conversations from regulated call centers.
Guava co-founder and CEO Omri Dahan said enterprises frequently encounter problems when moving voice AI systems from demonstrations into production environments.
He identified reliability during complex, high-volume customer calls as a major obstacle to broader adoption.
Dahan previously spent more than a decade as chief revenue officer at payments technology company Marqeta, helping expand the business before its public-market listing.
He established Guava with Gridspace technical co-founders Anthony Scodary and Nico Benitez, combining the research team’s voice AI technology with a dedicated commercial platform.
Acrew Capital Founding Partner Lauren Kolodny cited the combination of the founders’ technical experience and Dahan’s business development background as a central factor behind the investment.
Daytona combines speech recognition, AI models and voice synthesis
Unlike platforms that integrate speech recognition, language processing and voice generation technologies from multiple providers, Guava develops and operates its own models across the entire voice interaction process.
Guava says this architecture allows it to optimize model performance, response times and reliability without depending on separate vendors’ infrastructure or software release schedules.
The integrated system combines:
- Speech recognition: Converts customer speech into text and interpretable inputs.
- Language models: Process requests and determine appropriate responses.
- Speech synthesis: Generates spoken responses during conversations.
- Conversation management: Coordinates interactions through completion.
Its Daytona model has ranked among the top 10 systems on Coval’s word error rate benchmark, frequently achieving first- or second-place positions. The company is also developing full-duplex voice technology, which would allow AI agents to listen and speak simultaneously.
Additional investment will support deployment on edge devices and within customers’ own infrastructure.
Guava gives developers code-level control over AI conversations
Guava’s software development kit (SDK) allows businesses to incorporate voice agents into applications using Python, TypeScript or Claude Code.
The platform emphasizes software-defined business rules rather than relying primarily on natural-language prompts to control agent behavior.
Developers can connect company knowledge bases, establish operational requirements and integrate agents into existing workflows.
Guava manages the underlying speech models and conversation infrastructure, while customers determine the agent’s intended tasks and the information it can access.
The company says developers can create an initial working agent in less than five minutes using its free offering. Enterprise customers typically begin handling live calls in regulated environments within 14 days of implementation.
Anthony Scodary, Guava’s co-founder and co-head of engineering, said operating the entire voice AI stack allows the company to optimize its models together and maintain direct control over system performance.
Voice AI agent market projected to reach $43.8 bn by 2036
The investment comes as enterprises increasingly evaluate AI agents for customer service, financial operations and other telephone-based workflows.
According to Meticulous Research, the global voice AI agents market is projected to grow from $5.1 bn in 2026 to $43.8 bn by 2036.
Guava is positioning its platform for organizations seeking to replace or supplement conventional call center processes with automated voice interactions. Its initial focus includes enterprises and outsourcing providers operating in regulated industries, where voice agents must handle sensitive conversations within established business requirements.
Flourish Ventures co-founder and managing partner Emmalyn Shaw said consistent performance and developer control are necessary for wider enterprise deployment.
Guava will use the new funding to expand its commercial operations while continuing development of speech models, deployment tools and real-time conversational capabilities.
The company has not disclosed its valuation following the seed financing round.
