Traversal AI

Traversal is an AI site reliability engineering platform aimed at enterprise teams drowning in incidents and alerts. With this launch, Incident Workers move to general availability alongside the public beta of Alert Workers and a new Workers console for deploying and managing a fleet of agents across Slack and Microsoft Teams . The pitch to on-call engineers is straightforward: an Incident Worker joins an incident channel the moment something breaks, decides for itself when to engage, and owns the investigation end to end , while Alert Workers cut noise by learning which alerts are routine, deduping repetitive noise, and surfacing the handful that are new or meaningful, with the reasoning attached . The technical bet under the hood is causal reasoning over telemetry rather than a chat wrapper on top of dashboards. The Workers run on Traversal's Causal Search Engine and Production World Model, the same foundation behind the rest of the platform , which the company positions as a way to handle the messy, cross-service investigations that typically require a war room. This GA moment matters because it lands after real production use at scale. Since entering public beta on June 30, Workers have run across thousands of real incidents at some of the largest enterprises in the world, helping teams close incidents faster, pull fewer people into war rooms, and find the root cause before the full response team is paged , with American Express among the disclosed customers. Traversal was founded by AI researchers from MIT, Columbia, Berkeley, and Cornell , with Anish Agarwal as CEO and co-founder alongside Raj Agrawal, Ahmed Lone, and Raaz Dwivedi (a Cornell Tech assistant professor). The company is backed by Sequoia and Kleiner Perkins , and this release signals a shift from reactive AI copilots toward agents that decide on their own when to start investigating, a design choice founders and operators building agentic products will want to study closely.