InsightFinder raises $15M to help companies figure out where AI agents go wrong

The landscape of observability tools is shifting once more, as the growing demand for reliable tech systems continues to evolve. With the surge in AI agents across enterprises, a new category of workload has emerged, necessitating innovative monitoring solutions. InsightFinder AI, a startup built on 15 years of academic research, is addressing this challenge head-on. Since its inception in 2016, the company has leveraged machine learning to detect and resolve IT infrastructure issues. Now, it is tackling the pressing matter of AI model reliability with a comprehensive solution that encompasses detection, diagnosis, remediation, and prevention. Founded by Helen Gu, a computer science professor at North Carolina State University with experience at IBM and Google, InsightFinder recently completed a $15 million Series B funding round led by Yu Galaxy, as confirmed by TechCrunch. Gu emphasizes that the industry's primary challenge is not just pinpointing where AI models fail, but understanding how the entire tech stack functions with AI integrated into it. "To effectively diagnose AI model issues, it is crucial to monitor and analyze data, the model, and the infrastructure as a cohesive unit," Gu stated. "It’s often not just a model or data issue; it can also stem from infrastructure shortcomings." Gu illustrated this point with a real-world example involving a major U.S. credit card company that experienced model drift in its fraud detection system. InsightFinder's monitoring capabilities revealed that the drift was due to outdated cache data on certain server nodes. She noted a common misconception in the industry: that AI observability is confined to evaluating large language models during development phases. In reality, a robust AI observability platform should deliver comprehensive feedback across development, evaluation, and production stages. The company’s latest offering, Autonomous Reliability Insights, utilizes a blend of unsupervised machine learning, proprietary language models, predictive AI, and causal inference. This adaptable framework allows the system to process and analyze entire data streams, enabling effective correlation and validation to uncover root causes. As the observability market becomes increasingly competitive, InsightFinder competes with notable players like Grafana Labs, Fiddler, Datadog, Dynatrace, New Relic, and BigPanda, all of which are developing solutions for challenges introduced by AI tools. Despite the competition, Gu remains confident in InsightFinder’s unique capabilities and customizability, which she believes provide a significant competitive advantage. "We rarely lose clients to competitors; our strength lies in the insights we offer. Many data scientists grasp AI concepts yet overlook system intricacies, while site reliability engineers often lack AI understanding. This gap is where we excel." InsightFinder’s clientele includes prominent names such as UBS, NBCUniversal, Lenovo, Dell, Google Cloud, and Comcast. Gu credits their success to a decade spent comprehensively understanding the needs of large enterprise customers. "Collaborating with our Fortune 50 clients has been pivotal in refining our approach to deploying these models," she remarked, highlighting a significant partnership with Dell that has facilitated global deployments of their AI systems. The company’s revenue has reportedly surged over threefold in the past year. Interestingly, Gu noted that the company wasn't actively seeking to raise a Series B round; investors approached them following a notable seven-figure deal with a Fortune 50 client. The new funding will enable InsightFinder to expand its small team of fewer than 30, particularly in sales and marketing, and enhance its go-to-market strategy. To date, the company has successfully raised a total of $35 million.