Crux Labs raises over $6M to improve enterprise agents with human judgment

Crux Labs secured over $6 million in funding to enhance enterprise AI agents through the integration of human judgment. The company's innovative platform transforms enterprise judgment into continuous agent improvement, addressing the challenge of aligning AI agents with human expertise. Benchstrength and Misneach led the funding round, with Analog Ventures and Pareto Holdings joining in.

The scale of enterprise AI agents can vary from thousands to millions of user interactions daily. Although observability platforms facilitate human monitoring of agentic traces, they complicate the process of improving agent behavior at scale. Many teams resort to sampling only a fraction of traces for manual review, resulting in the majority of agent failures going unnoticed.

Julie Kheyfets, Crux Labs' co-founder and CEO, emphasized the difficulty of aligning AI agents with human judgment, stating that while AI agents excel at achieving goals, they lack the collective heuristics humans have accumulated over time. To bridge this gap, Crux Labs employs its own agents to capture expert judgment from employees and calibrate it across millions of interactions, enabling workers to amplify their understanding without analyzing a limited sample.

The platform has demonstrated significant traction among high-growth startups and Fortune 500 companies in various complex industries, including healthcare, cybersecurity, and financial services. Co-founder and Chief Technology Officer Krishna Kulkarni highlighted the need for a bridge between engineering teams, which build AI agents, and subject matter experts who define success criteria.