What Coval does

Coval is a San Francisco-based voice AI evaluation platform founded in 2024 by Brooke Hopkins, providing simulation, observability, and human-in-the-loop review infrastructure for enterprises deploying autonomous voice agents. The platform operates as a continuous quality loop across the agent lifecycle through three integrated modules — Simulate (pre-deployment stress-testing with thousands of realistic scenarios including accents, IVRs, DTMF inputs, and background noise), Observe (real-time production scoring across latency, escalation, identity verification, sentiment, and compliance metrics), and Human Review (smart-sampling of failures to human reviewers whose feedback retrains the AI judge). Coval is API-first, exposing a REST API at api.coval.dev/v1 alongside TypeScript and Python SDKs, a CLI for CI/CD integration, an MCP server for native access from AI coding tools (Claude Code, Cursor, Windsurf, Codex), and OpenAPI specifications at public endpoints.

The company serves 60+ enterprise customers including Zoom, Deepgram, Perplexity, ServiceNow, Chime, Hippocratic AI, GEICO, Toast, StubHub, and Upstart, spanning communications, fintech, healthcare, insurance, restaurant tech, ticketing, and AI search. Customer-reported outcomes include up to 30x reduction in manual QA processes, up to 10x faster agent deployment times, and a 217% improvement in agent accuracy within seven days of onboarding. Coval targets four primary use cases: agent platforms proving production-readiness to their customers, in-house engineering teams future-proofing their stacks, behavior-focused evaluators testing for identity/escalation/hallucination failures, and procurement-led vendor bakeoffs benchmarking multiple voice AI providers on identical scenarios.

Revenue derives from SaaS subscriptions with a usage-based component tied to evaluation volume; distribution combines product-led growth (free voice agent audit, free trial, self-serve CLI/API onboarding) with a direct enterprise sales motion supported by an expanding solutions engineering team. The company has raised $31 million in total funding since its 2024 launch, including a $28 million Series A closed in June 2026 led by Norwest with participation from Base10 Partners, Twilio Ventures, and Y Combinator. Compliance initiatives include SOC 2 Type II (in progress), GDPR, and HIPAA, enabling deployment in regulated industries.

Name

Coval

Legal name

Coval, Inc.

Website

https://coval.dev

Company type

Private

Founded year

2024

Operating status

Operating

Headcount range

11–50 employees

Short description

Coval is an AI-native evaluation platform for autonomous voice agents, providing pre-deployment simulation, real-time production monitoring, and human-in-the-loop review. Its API-first platform serves 60+ enterprise customers including Zoom, Deepgram, Perplexity, and ServiceNow, processing 3.1 million evaluation metrics weekly.

Ownership category

akta.pro rank

Product category

AI Evaluation and Testing Infrastructure

NAICS

Computer Systems Design and Related Services (54151)

SIC

Services-Prepackaged Software (7372)

akta.pro primary industry

Conversational AI Platforms (chat/voice bots, orchestration) (HDAAAFAC)

akta.pro secondary industries

Contact Center Speech AI (agent assist, IVR automation) (HDAAAFAL), Voice Assistants & Voice UI (VUI) (HDAAAFAD)

Keywords

• Voice AI evaluation
• AI agent simulation
• Production observability
• Automated quality assurance
• Human-in-the-loop review

Headquarters

HQ city

San Francisco

HQ country

United States

HQ region

North America

Offices1 record

Markets served

GTM type

B2B

Offering type

Software

Cost components

Personnel, Technology or R&D, Marketing or Sales, Infrastructure, Operations

Revenue model

1. SaaS Subscription: Coval operates as a SaaS platform with subscription-based pricing. The platform is described as deployment-ready infrastructure with tiered plans available via a pricing page.
2. Usage-based Evaluation: The platform processes tens of millions of evaluations for enterprises and runs 3.1M evaluation metrics weekly, suggesting a usage-based component tied to evaluation volume.

Pricing tiers

Go-to-market motion2 records

Distribution channels4 records

Marketing channels9 records

Core offering

Coval provides a unified deployment-readiness platform for autonomous voice AI agents, consisting of three integrated modules: Simulate (pre-deployment testing of thousands of realistic conversations), Observe (real-time production call scoring and regression detection), and Human Review (human-in-the-loop feedback that retrains the AI judge). The platform is delivered via a REST API, CLI, TypeScript and Python SDKs, MCP server, and public benchmarks portal.

Product overview

Coval is a unified deployment-readiness platform for voice AI agents, consisting of three integrated modules: Simulate (pre-deployment testing), Observe (production monitoring), and Human Review (human-in-the-loop improvement). The platform enables enterprises to test voice agents before launch, monitor them in production, and continuously improve evaluation quality through human feedback. Customers can benchmark agents on equal footing, stress-test scenarios via CLI or API, and automate QA processes across the full agent lifecycle.

Differentiator

Problem solved

Functional benefit

Products and services

• Simulate Pre-deployment testing module that runs thousands of realistic conversations with realistic voice inputs, IVRs, accents, background noise, and edge cases to stress-test voice AI agents before launch. Runs via CLI in under 15 minutes.
• Observe Production monitoring module that scores every live voice AI call in real time, surfacing regressions, latency issues, escalation failures, and compliance failures as they occur.
• Human Review Human-in-the-loop review module that routes evaluation failures to human reviewers whose feedback retrains the AI judge and sharpens evaluation metrics. Includes AI verdicts with one-click human override.

Quantifiable outcome

• Reduce manual QA by up to 30x
• +4 more outcomes

Named customers10 records

Segments11 records

Ideal customer profiles3 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

Integration6 records

AI capability6 records

Feature7 records

Partnerships

One partnership is on record.

• CB Insights (Future Tech Hotshots 2025)minorOthers · 1 October 2025Coval was identified by CB Insights as one of 45 emerging tech startups expected to have significant impact over the next 5-10 years in AI infrastructure.

Scale indicators10 records

Recent moves6 records

Expansion highlights7 records

Direct peers

• Braintrust: Braintrust is an AI evaluation and observability platform for production LLM and AI applications. It directly competes with Coval in providing end-to-end evaluation infrastructure, though Coval is specialized in voice modalities while Braintrust started broader.
• LangSmith (LangChain): LangSmith is LangChain's evaluation, monitoring, and debugging platform for LLM-powered applications. It is a direct competitor in the AI agent evaluation space, with growing coverage of agent workflows that overlaps with Coval's continuous quality loop.
• Arize AI: Arize AI provides ML observability and LLM evaluation infrastructure for production AI systems. It competes with Coval's Observe module and is expanding into agent evaluation workflows relevant to voice AI deployments.
• Galileo (Galileo Technologies): Galileo offers an evaluation and observability platform for generative AI and LLM applications. It competes with Coval across agent evaluation, scoring, and human-in-the-loop workflows, with comparable enterprise go-to-market.
• Confident AI (DeepEval): Confident AI builds DeepEval, an open-source LLM evaluation framework, and a commercial evaluation platform. It targets similar AI builder personas to Coval with comparable testing and benchmarking functionality for AI applications.
• WhyLabs: WhyLabs provides AI observability and monitoring for ML and LLM systems in production. It competes with Coval's Observe module in real-time production monitoring and quality assurance for deployed AI agents.

Broad incumbents

• Observe.AI: Observe.AI is a contact center AI platform providing quality monitoring, agent evaluation, and conversation intelligence. It overlaps with Coval in voice AI quality scoring and contact center deployments, but serves a broader contact center suite rather than voice agent evaluation specifically.
• CallMiner: CallMiner is an established conversation analytics platform for contact centers, providing speech analytics, quality monitoring, and compliance scoring. It is a broader incumbent in voice analytics that overlaps with Coval's evaluation and compliance scoring use cases.
• Scale AI: Scale AI provides data infrastructure, evaluation, and fine-tuning services for AI applications. It is a broader incumbent in AI evaluation and testing that could expand deeper into voice agent evaluation, competing with Coval on enterprise AI testing.
• Hugging Face: Hugging Face offers open evaluation tools (e.g., evaluate library) and a broad AI platform. As an open-source and platform incumbent, it overlaps thematically with Coval's evaluation infrastructure for AI models and agents, though with a much broader scope.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks5 records

Key highlights7 records

Customer concentration

• Xx.com/covaldev
• LinkedInlinkedin.com/company/covaldev

Compliance3 records

Revenue estimate

Valuation estimate

Number of profiles

Profiles2 records

Funding overview

Funding rounds4 records

Investors11 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Coval