Trusted by Zoom, Deepgram, and different Fortune 500 firms, voice AI testing platform Coval brings enterprise-grade analysis infrastructure to autonomous voice brokers
Coval, the analysis platform for voice AI, introduced a $28 million Collection A spherical of financing led by Norwest with participation from Base10 Companions, Twilio Ventures and Y Combinator. This brings the overall capital raised to $31 million since its launch in 2024.
Coval is the main simulation, observability and labelling platform for AI voice and chat brokers that enables enterprises to scale voice and chat AI brokers. As Fortune 500 firms look to deploy voice AI and discover the way it can change the paradigm for his or her companies, Coval offers them the infrastructure to do it reliably and with confidence. The brand new funding will allow Coval to handle the rising reliability and compliance challenges enterprises face as they deploy autonomous voice brokers. Coval runs tens of hundreds of thousands of evaluations, and with this new capital, the corporate will increase its gross sales and options engineering groups to scale and meet this rising demand. Coval can even advance product capabilities together with deeper simulation, new integrations, and enhanced human evaluate and monitoring options.
“Each firm goes to have a voice agent identical to they’ve a cellular app or an internet app. However right now, most enterprises don’t have the infrastructure to deploy these programs with confidence,” stated Brooke Hopkins, founder and CEO of Coval. “Coval offers groups the power to simulate, monitor and constantly enhance voice brokers, to allow them to transfer from experimentation to dependable manufacturing at scale.”
Coval’s Full‑Stack Platform for the Whole Voice Agent Lifecycle
The voice recognition market is rising quickly, with greater than $7 billion invested in voice AI within the first quarter of 2026 alone and expectations that it’ll attain greater than $20 billion by 2031.
Enterprises are quickly adopting voice AI for customer support, gross sales, monetary providers and healthcare, however most nonetheless depend on guide QA processes that break below actual‑world complexity and scale. Coval’s platform empowers enterprises with a complete testing and monitoring infrastructure for voice AI.
“Voice goes to be the primary interface for the way people work together with AI, and that shift creates a completely new infrastructure layer for enterprises,” stated Scott Beechuk, associate at Norwest. “Along with her deep expertise constructing analysis programs for autonomous applied sciences at Waymo, Brooke is uniquely positioned to steer Coval in defining how firms deploy and scale voice brokers reliably. She helped show self-driving vehicles may work, and now she’s tackling voice AI.”
In contrast to different analysis instruments which might be constructed for builders, Coval is constructed to assist groups throughout operations, QA, engineering and product collaborate and scale autonomous brokers. Coval affords a full‑stack platform spanning your complete voice agent lifecycle, from pre‑deployment simulation and stay manufacturing monitoring to human evaluate and structured analysis. Its structure is objective‑constructed for voice, audio processing and high quality evaluation, telephony latency, transcription error evaluation and analysis of agent workflows. The corporate’s autonomous‑programs roots additionally set it aside, making use of the simulation‑first self-discipline Hopkins developed at Waymo to carry manufacturing‑grade rigor to voice AI. Just like self-driving vehicles, voice brokers navigate the world autonomously by operating fashions in parallel to hear (transcription), cause (LLMs) and converse (text-to-speech) in the identical approach that self-driving vehicles use notion, planning and controls to navigate the world. Testing between the 2 programs have comparable parallels – simulation is essential for testing voice AI functions.
Additionally Learn: AiThority Interview with Matej Bukovinski, Chief Expertise Officer at Nutrient
“Twilio’s open, versatile, and model-agnostic infrastructure for the agentic period is driving the shift towards human-like voice AI agent experiences changing into the norm in buyer engagement,” stated Andy O’Dower, VP, Discipline CTO at Twilio. “Belief is essential to scaling these experiences, and our funding in Coval displays our conviction that complete analysis and testing instruments, mixed with a powerful observability and reliability layer, are foundational to sustaining momentum in right now’s voice AI renaissance.”
Coval automates voice AI agent testing to constantly enhance analysis accuracy. The platform allows enterprises to run probabilistic evaluations throughout hundreds of thousands of voice interactions. Firms like Zoom, together with greater than 60 extra enterprises, depend on Coval to scale back guide QA processes by as much as 30x and improve voice agent deployment occasions by as much as 10x.
Firms Like Zoom and DeepgramTrust Coval to Ship Higher Voice AI Experiences
“Reliability and observability are a high precedence for us at Zoom as voice AI strikes into customer-facing manufacturing environments,” stated Ram Rajagopalan, Head of Product – CX AI at Zoom. “Coval offers Zoom’s prospects the power to judge conversations systematically at scale, determine edge instances earlier than they impression customers, and transfer considerably quicker with confidence.”
“Voice brokers introduce a brand new degree of complexity in comparison with conventional software program testing,” stated Anoop Dawar, COO at Deepgram. “Brooke has constructed Coval right into a core a part of the trendy enterprise’s analysis stack by bettering reliability earlier than scaled deployment. For any critical enterprise deployment, that is now not a nice-to-have. At Deepgram, we energy the voice AI infrastructure groups construct on, however due to our partnership with Coval, enterprises can relaxation assured it’s working correctly.”
