- Crux Labs has raised over $6 million, with Benchstrength and Misneach leading the round.
- The founders have experience at Tractable, Block Renovation, Meta and Checkr.
- Crux plans to use the funds for R&D and to accelerate enterprise deployments.
Most companies still manually review only 1% of their AI agent activity. Crux Labs, a New York-based startup, uses software to review the rest and is exiting stealth with over $6 million raised.
Benchstrength and Misneach led the oversubscribed round, with Analog Ventures and Pareto Holdings on board.
According to Crux Labs, its AI agents can connect to customers’ tools, learn from the company’s experts, and use that knowledge to generate custom tests. They then run those tests on all production traffic, generating alerts and enabling rapid fixes for the live agents.
In domains like healthcare and financial services, this means applying clinicians’ or compliance officers’ standards to every interaction.
Moving from insurance claims and renovations to auditing AI agents
Crux’s founders met at Princeton. “Aligning AI agents to human judgment is one of the biggest challenges of our time,” says Julie Kheyfets, co-founder and CEO.
Kheyfets previously led Tractable’s North America business, a London-based company that uses AI to assess accident damage from a photo and raised $65m in its Series E round in 2023. She then joined Block Renovation, a Brooklyn-based renovation marketplace, and became CEO in January 2025.
In both roles, the software made decisions that would normally fall to an expert. Kheyfets now sees the same gap with AI agents: “Agents are incredible at attaining goals, but they lack access to the years of heuristics that we as humans have collected in our heads. Crux helps every enterprise to distill this judgment and to use it to rapidly improve their agents.”
Co-founder and CTO Krishna Kulkarni led brand engineering at Meta, built AI products at Checkr and took Healthify from zero to 20m patients.
“Engineering teams know how to build AI agents, and subject matter experts know what success looks like, but enterprises lack the bridge between them. Crux operationalizes domain expertise into continuous agent improvement so engineers can build truly business-aligned agents,” Kulkarni said.
A crowded market with ongoing buyer interest
Crux is entering a space that has seen significant investment and recent M&A. In February, Braintrust raised $80 million at an $800 million valuation. Cisco has announced its intention to acquire Galileo. ClickHouse acquired Langfuse in January. Patronus AI raised $50 million in a Series B, and Judgment Labs closed a $32 million round led by Lightspeed.
Unlike rivals, Crux adjusts its scoring to align with each customer’s own experts. It works with several Fortune 500 companies and fast-growing startups in the health, cyber, and financial services space but has not disclosed any customer names, revenue, or results.
Demand for AI observability is up: the market is expected to grow from $1.24 billion in 2025 to $12.75 billion by 2032. More work is being handed to agents, as evidenced by Hyphenate raising $45 million for their finance agents and Fragment raising $56 million for their procurement exceptions
Where the funding comes from and how it will be used
Crux plans to use the funds for R&D and to accelerate enterprise deployments. They are hiring researchers and engineers and are being led by a female co-founder and CEO.
“The companies that win with AI will have agents that reflect the judgment of their best people. Crux makes this possible across every customer interaction, and Julie and Krishna are uniquely suited to build it,” says John Monagle, general partner at Benchstrength.
Investors and buyers largely agree that agent oversight is important. The real issue is whether a young company can actually own the process of turning expert intuition into tests, or whether it will simply become another feature on a bigger platform.
