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

The role of observation instruments has evolved again. As the market for solutions to ensure the reliability of engineering systems has grown over the years, the focus has steadily shifted from ‘tracking everything’ to ‘managing complexity and costs’. Meanwhile, the rapid influx and adoption of AI agents within enterprises has only added a whole new category of workload to consider.

InsightFinder AIa startup based on 15 years of academic research, is no stranger to this problem.

The company has been using machine learning to monitor, identify and proactively resolve IT infrastructure issues since 2016, and is now addressing the current AI model reliability problem with an AI agent solution that can do everything from detection and diagnosis to remediation and prevention.

The company, founded by CEO Helen Gu, a computer science professor at North Carolina State University who previously worked at IBM and Google, recently raised $15 million in a Series B round led by Yu Galaxy, TechCrunch has exclusively learned.

According to Gu, the biggest problem facing the industry today is not just monitoring and diagnosing where AI models go wrong; it diagnoses how the entire tech stack works now that AI is part of it.

“To diagnose these AI model problems, you have to actually monitor and analyze the data, the model, and the infrastructure together,” Gu told TechCrunch. “It’s not always a model problem or a data problem; it’s a combination. Sometimes it’s just your infrastructure.”

Gu explained what that looks like in real life with an anecdote: One of his customers, a major U.S. credit card company, noticed that one of its fraud detection models was drifting. Because InsightFinder monitored the company’s entire infrastructure, it was able to determine that the model discrepancy was caused by outdated cache in some server nodes.

“The biggest misconception is that AI observability is limited to LLM evaluation during the development and testing phases. On the contrary, a good AI observability platform should provide end-to-end feedback loop support that spans the development, evaluation and production phases,” she said.

InsightFinder’s latest product, called Autonomous Reliability Insights, can do all this by using a combination of unsupervised machine learning, proprietary large and small modeling language models, predictive AI, and causal inference. This base layer, according to Gu, is data-agnostic, allowing the system to ingest and analyze entire data streams to collect signals that can then be correlated and cross-validated to arrive at a root cause.

Now the observation room is full of candidates for some of the new market opened up by the influx of AI tools. Nearly a decade into its journey, InsightFinder is competing against companies like Grafana Labs, Fiddler, Datadog, Dynatrace, New Relic, and BigPanda, all of which are building capabilities to address the emerging problems of AI tools.

But Gu is not impressed. Rather, she argues that the InsightFinder’s expertise, experience, and adaptability act as a sufficient moat. “We actually rarely lose [customers] for everyone so far […] It’s about the insights, right? The problem is that many data scientists understand AI, but not the system. And lots of SRE [site reliability engineering] developers understand the system, but not the AI […] They don’t look at it and don’t understand the intrinsic relationships.”

Today, InsightFinder’s customer base includes UBS, NBCUniversal, Lenovo, Dell, Google Cloud and Comcast, and Gu attributes its success to 10 years of work understanding what large enterprise customers need.

“The bottom line is that we need to work with our Fortune 50 customers to improve and understand the business environment requirements to deploy these types of models,” she says. “We’ve been working with Dell to deploy our AI systems around the world to some of the largest customers we have. This isn’t something where you can take basic AI and just use the machine data.”

Gu said the company’s revenue stream is “strong” and has grown “more than threefold” in the past year. In fact, she says the company had no intention of raising this Series B at all, and that investors approached the company after the company closed a seven-figure deal with a Fortune 50 company within three months.

InsightFinder will use the new capital to recruit its first sales and marketing employees, expand its team of under 30 people and invest in the go-to-market movement. The company has raised a total of $35 million to date.