Crux Labs has emerged from stealth with an oversubscribed funding round of more than $6 million, the company announced on October 29, 2026. The round was co-led by Benchstrength and Misneach, with participation from Analog Ventures and Pareto Holdings, and will fund expanded research and engineering plus faster deployments with enterprise customers.
Crux tackles a problem that has grown alongside the agent boom: companies are deploying AI agents at scale without a reliable way to measure or improve them. Many teams manually review just 1% of agent traces, missing failures that hide in the long tail of millions of interactions. Existing observability tools make traces easier to inspect, the company argues, but do not systematically improve performance.
The Crux platform captures expert judgment inside a company’s workflows and calibrates custom evaluations against it. Those evaluations then monitor 100% of production traffic in real time and feed continuous improvements directly back into production. The pitch is to give every enterprise the capabilities of a dedicated evaluation science team, freeing engineers to ship features instead of triaging agent failures.
The company reports traction with startups and Fortune 500 enterprises across healthcare, cybersecurity, and financial services. Cofounders Julie Kheyfets and Krishna Kulkarni met at Princeton; Kheyfets previously led North America for Tractable AI, while Kulkarni led brand engineering at Meta and scaled Healthify to 20 million patients.
As agents move from demos into regulated, high-stakes work, evaluation is becoming its own category. Our guide to what AI agents are explains the basics; Crux wants to own the next step, proving those agents behave the way a company’s best experts would. If it works, the winners in enterprise AI may be decided less by who builds the flashiest agent and more by who can show it works reliably.
Source: Morningstar / PR Newswire
