InsightFinder Raises $15M to Debug AI Agents

The landscape of enterprise technology is being reshaped by the widespread deployment of AI agents, creating a critical new demand for sophisticated observability tools. This shift moves beyond simply tracking system metrics toward actively managing complexity and cost, with a specific focus on ensuring the reliability of AI-powered workflows. A startup born from over a decade of academic research is now securing significant funding to tackle this very challenge.

InsightFinder AI, which has applied machine learning to IT infrastructure monitoring since 2016, is directing its expertise toward the reliability of AI models. The company recently closed a $15 million Series B funding round led by Yu Galaxy. Founded by North Carolina State University professor and former IBM and Google engineer Helen Gu, the firm is addressing what Gu identifies as the core industry problem: diagnosing failures across the entire integrated tech stack, not just within an isolated AI model.

“To properly diagnose these AI model problems, you must monitor and analyze the data, the model, and the underlying infrastructure together,” Gu explained. “The issue isn’t always purely the model or the data, it’s often a combination. In many cases, the root cause is simply your infrastructure.” She illustrated this with a case involving a major U. S. credit card company, where InsightFinder’s platform detected model drift in a fraud detection system. By monitoring the full infrastructure, the tool identified the cause as outdated cache residing in specific server nodes, a problem invisible to siloed monitoring.

Gu emphasizes that effective AI observability must extend far beyond evaluating large language models during development. “A sound AI observability platform should provide end-to-end feedback loop support covering development, evaluation, and production stages,” she stated. The company’s newest product, Autonomous Reliability Insights, aims to deliver this by leveraging unsupervised machine learning, proprietary language models, predictive AI, and causal inference. This data-agnostic base layer ingests complete data streams, correlating signals to pinpoint root causes.

The market for these solutions is becoming increasingly competitive, with established players like Grafana Labs, Datadog, and Dynatrace all enhancing their offerings for the AI era. Gu, however, believes InsightFinder’s deep expertise and customizability provide a strong defense. The company’s unique value, she argues, lies in bridging a significant knowledge gap. “Many data scientists understand AI but not the system, while many site reliability engineers understand the system but not the AI. They often miss the intrinsic relationships between them,” Gu noted, adding that this holistic insight has helped the firm retain its client base.

That client roster includes prominent names such as UBS, NBCUniversal, Lenovo, Dell, Google Cloud, and Comcast. Gu credits this success to a decade of refining the platform to meet the exacting demands of large enterprises. “It came down to working with our Fortune 50 customers to polish and understand the specific requirements for deploying these models in complex enterprise environments,” she said. “Deploying AI systems globally for some of the largest customers is not something you can do by just applying a foundational AI model to machine data.”

The company reports a “strong” revenue stream that grew more than threefold last year. Gu revealed that the Series B round was investor-led, initiated after InsightFinder secured a seven-figure contract with a Fortune 50 company within a three-month period. The new capital will be used for the company’s first dedicated sales and marketing hires, expanding its sub-30 person team, and accelerating its go-to-market strategy. Total funding to date now stands at $35 million.