Crux Labs exits stealth with $6M seed to audit AI agents against expert judgment
What's the deal? CruxDealroom has a profile for this one. Try Dealroom → Labs, a New York-based startup, has raised over $6 million in seed funding to check whether enterprise AI agents meet the standards of a company's own experts. BenchstrengthDealroom has a profile for this one. Try Dealroom → and MisneachDealroom has a profile for this one. Try Dealroom → led the oversubscribed round, with Analog VenturesDealroom has a profile for this one. Try Dealroom → and Pareto HoldingsDealroom has a profile for this one. Try Dealroom → participating.
Why now? Most companies manually review only about 1% of their AI agent activity, meaning the vast majority of failures go unnoticed. As enterprises hand agents thousands to millions of interactions a day, that gap widens.
What's the endgame? Crux connects to a customer's tools, learns from its experts, and generates custom tests that run across all production traffic, flagging problems for rapid fixes. In healthcare or financial services, that means applying clinicians' or compliance officers' standards to every interaction.
Co-founder and chief executive officer Julie Kheyfets previously led Tractable's North America business and became CEO of Block Renovation in January 2025. Co-founder and chief technology officer Krishna Kulkarni led brand engineering at Meta and built AI products at Checkr.
"Agents are incredible at attaining goals, but they lack access to the years of heuristics that we as humans have collected in our heads," Kheyfets said. The company plans to spend the money on research and development and on accelerating enterprise deployments.
What could go wrong? Crux says it works with several Fortune 500 companies and fast-growing startups in health, cyber, and financial services, but has not disclosed any customer names, revenue, or results. Whether a young company can own the process of turning expert intuition into tests — or becomes a feature on a bigger platform — remains open.
The signal: Crux is entering a crowded market. Evaluation and observability startups raised $272 million through September 2026, up from $166 million a year earlier, and Dynatrace agreed to acquire Arize in a reported $915 million deal. Crux's round is modest by comparison, but it targets the gap between simply watching agents and deciding whether their work meets a business's standards.
Read more: Tech Funding News
Image credit: Generated with Gemini
