Traversal
Traversal builds AI agents that triage alerts, find incident root cause, and block risky infrastructure changes.
Founded
2024Traversal builds AI agents that triage alerts, find incident root cause, and block risky infrastructure changes.
The founding team pairs causal machine learning research from Columbia and Cornell Tech with production and trading-systems engineering, which shapes a diagnosis method built on causal inference rather than alert correlation alone.
American Express acts as both customer and strategic investor, giving the company a reference deployment inside a regulated financial institution. Publishing head-to-head comparisons against Resolve AI and Datadog also positions the product against agent startups and observability incumbents at the same time.
Source: traversal.com
Enterprise reliability budgets sit today with observability platforms that report symptoms, while the emerging category sells diagnosis and remediation. Traversal, Resolve AI, and Cleric each pursue that shift with their own agents aimed at the same platform engineering buyer.
Incumbents such as Datadog can attach agent features to contracts customers already hold, so displacement depends on proving accuracy in live production rather than on feature count. Regulated buyers add a second requirement, since an agent that touches infrastructure needs auditable controls.
Source: devops.com
Traversal Workers operate on a production world model and a causal search engine that keep an updated map of a customer's services, code, and infrastructure dependencies. The agents triage alerts without prompting, run root-cause analysis, and remediate failure conditions they recognize.
The same platform screens pending change: it previews the blast radius of a proposed change, can block a risky pull request before deployment, and answers production support questions from engineers. Named customers include DigitalOcean, Cloudways, PepsiCo, and American Express.
Source: traversal.com