Downtime is a crisis: one that costs global businesses an estimated $400 billion a year and can drain up to $1.9 million per hour during major incidents. For engineering teams, production failures often feel like cardiac arrests. As Anish Agarwal, co-founder and CEO of Traversal, put it: “An engineering team has two heart attacks a week.”
Fresh out of stealth with $48 million in combined seed and Series A funding from Sequoia Capital, Kleiner Perkins, NFDG, and Hanabi, Traversal is betting that artificial intelligence (specifically, a hybrid of causal machine learning and agentic AI) can finally crack one of the oldest, most persistent problems in software: root cause analysis. Their goal is to build an AI Site Reliability Engineer (SRE) for enterprise.
Despite decades of progress in observability tools (think Datadog, Splunk, New Relic) engineers still end up “dashboard dumpster diving,” frantically sifting through petabytes of logs and metrics to pinpoint why a system crashed. Agarwal, along with co-founders Raaz Dwivedi, Ahmed Lone, and Raj Agrawal, saw this problem play out repeatedly in both academic and high-stakes production environments.
The team’s backgrounds span academia and industry: Columbia and Cornell faculty, a former Citadel Securities quant trader, and AI researchers deeply embedded in the science of causality.
How Traversal Works: An Agentic AI Approach to Observability
Traditional observability platforms that intimate engineers with visualizations and alerts, Traversal takes an agentic approach.
Its AI agent is designed to autonomously troubleshoot, remediate, and prevent complex production incidents. The system operates by “traversing” massive datasets and orchestrating calls to existing tools. A standout is its ability to infer causality. Instead of surfacing symptoms, Traversal works to identify the true root causes, often in real time.
Traversal is positioning itself as both a standalone platform and an intelligence layer that enhances existing observability stacks. Its business model targets enterprise customers with high reliability demands: cloud infrastructure providers, fintech firms, payment processors, and large-scale SaaS companies. In these sectors, even a few minutes of downtime can ripple out to customer attrition, reputational damage, and lost revenue.
The platform is already live in high-stakes production environments at companies including Eventbrite, Cloudways, American Express, and several undisclosed financial institutions. These are daily-use deployments where Traversal serves as the first responder to software incidents.
What makes this go-to-market strategy especially sticky is that the product tends to “self-select” for long-term customers. Once integrated, it provides immediate ROI in terms of reduced downtime and engineer productivity. That makes it hard to rip out and easy to expand.
Riding the AI Agent Wave
Traversal is tapping into one of the most important currents in today’s AI landscape: agentic AI. As software systems grow more complex, partly because of the rise of AI code generation, humans are increasingly ill-equipped to manage them on their own. LLMs like GitHub Copilot are accelerating software creation, but they’re also introducing new, opaque layers of logic that make failures harder to trace.
Kleiner Perkins partner Mamoon Hamid believes that observability needs a new playbook entirely. “With more code created by AI, there is more surface area to troubleshoot,” he said. “There is a need for AI to autonomously troubleshoot, mediate and even prevent complex incidents at scale – self-healing codegen.”
This positions Traversal not as a replacement for existing tools like Splunk or Datadog, but as the intelligence layer that turns observability into action. Its system doesn’t require everything to be pre-defined in a runbook since it can investigate unknown unknowns using parallel, probabilistic search techniques born out of research in statistical causality.
Investor Backing and a Long-Term Bet
Traversal’s funding speaks to the level of conviction among top-tier investors. Sequoia led the seed round, while Kleiner Perkins led the Series A. Nat Friedman and Daniel Gross’s NFDG and Hanabi Ventures also joined in, alongside infrastructure-focused angels and institutional backers. What drew them in was the team’s technical depth and clarity of vision.
“Troubleshooting is one of the most painful and expensive workflows in software,” said Sequoia partner Bogomil Balkansky. “It’s why we believe that observability, infused with AI, is the next frontier.”
What makes Traversal stand out in an increasingly crowded AI startup field is its research-first foundation. The company remains deeply embedded in AI science, with plans to grow into a kind of “agent lab” for enterprise software. That means pushing the limits of what agents can do, not just in theory, but in real-world production systems.
Traversal is not the only company trying to bring AI into observability, but it may be among the most academically grounded and operationally focused. As Agarwal described, today’s SREs spend their days dealing with “heart attacks” and “chronic conditions.” Traversal wants to help them shift toward planning, architecture, and growth.
