Coval Raises $28M Series A to Make Voice AI Reliable in Production

At Waymo, I spent my days on a question that sounds simple and is anything but: how do you know a self-driving car is safe enough to put on a public road? You can’t drive it around the block a few times and call it good. The only honest answer is infrastructure. Millions of simulated miles. Regression suites that catch the one behavior that quietly changed. Production metrics that tell you the truth after the car is out in the world. That is how a probabilistic system earns trust.

When I left Waymo, I realized the same question was about to land on every company in the world. Every enterprise is going to have a voice agent, the same way every company today has a website and a mobile app. And almost none of them have the infrastructure to deploy one with confidence.

That is why I started Coval. And today, I am thrilled to announce that we have raised a $28 million Series A, led by Norwest, with participation from Base10 Partners, Twilio Ventures, Y Combinator, MaC Ventures, and Swift Ventures. This brings our total raised to $31 million since we founded the company in 2024.

Why we’re building Coval

Voice AI is not a “someday technology.” Enterprises are already putting agents on the phone for things like customer support, scheduling, collections, and patient intake, handling millions of real conversations a day where a wrong answer can cost money, trust, or compliance. But the industry has raced ahead of its own reliability: most of those agents demo beautifully and then stumble the moment real traffic hits them. Roughly 95% of voice agents work in a demo, and only about 62% survive their first week live, brought down by the accents, interruptions, background noise a quiet room never shows in demo. It’s not a model-quality problem, it’s an infrastructure problem, and it’s playing out right now at companies your customers call every day. Closing that gap is the entire reason Coval exists: making sure agents handling real conversations work consistently and compliantly, call after call.

Why now

More than $7 billion went into voice AI in the first quarter of 2026 alone, and the market is expected to pass $20 billion by 2031. Every dollar of that builds another agent that someone then has to deploy reliably. That’s the layer we live in.

I could not have asked for better partners to build this next chapter with. Scott Beechuk, the partner leading our round at Norwest, put it this way: “Voice is going to be the number one interface for how humans interact with AI, and that shift creates an entirely new infrastructure layer for enterprises. … She helped prove self-driving cars could work, and now she’s tackling voice AI.”

What’s next

This funding goes to two main places: the people who help enterprises reach production, and the product underneath them. We’re growing our sales and solutions engineering teams, because getting a voice agent to real reliability is as much a methodology problem as a tooling one. And we’re going deeper on the product: richer simulation across more accents, audio conditions, and adversarial callers; new integrations so Coval drops cleanly into whatever stack you already run; and stronger human review and monitoring that close the loop from production right back into your tests.

To our customers, our team, and our new partners: thank you for seeing the same future we do. The agents are coming whether the infrastructure is ready or not, and the companies that win this era will be the ones that can deploy with confidence. The hardest, most interesting version of this problem is still ahead of us, and we’re growing the team that gets to solve it. If making voice AI reliable sounds like the work you want to do, come build it with us.

Brooke Hopkins Founder and CEO, Coval