Coval raises $28M for voice AI evaluation
Coval raised a $28M Series A to build evaluation tools for voice AI agents.
Voice AI is moving into real customer-facing workflows, but reliability is still the hard part. Coval is building the testing layer companies need before trusting voice agents in production.
What happened
Coval, a San Francisco-based voice AI evaluation platform, raised a $28M Series A led by Norwest.
The company helps teams evaluate voice AI systems before they are deployed into workflows such as customer support, sales, healthcare intake and other spoken-interaction use cases.
Why it matters
This is a strong enterprise AI infrastructure signal.
Voice agents need to be tested differently from text agents because they deal with timing, tone, interruptions, accents and real-time customer conversations. Poor performance can quickly damage customer experience.
The bigger picture
As AI agents move from chat boxes into calls and live interactions, evaluation will become core infrastructure. The next wave of voice AI adoption may depend on tools that prove these systems are accurate, safe and reliable enough for real users.