Traversal
by Traversal
Enterprise AI SRE for alert triage and root cause analysis across existing observability tools
Traversal is an AI site reliability engineering platform for large engineering organizations running complex production systems. It triages alerts, traces incidents to a root cause across services and code changes, and helps drive remediation, working on top of the observability tools a team already pays for.
Traversal is an AI SRE platform founded in 2023 by Anish Agarwal, Raaz Dwivedi, Raj Agrawal and Ahmed Lone, three of whom came from academic research in causal machine learning and reinforcement learning. The product rests on two components. The Production World Model is a continuously updated representation of a production system built from telemetry, code, deployment history and team knowledge. The Causal Search Engine, introduced in March 2026, tests many hypotheses in parallel and follows causal links across systems until it reaches an evidence-backed root cause. Traversal Workers are long-running agents that start when an alert or incident fires, and Traversal MCP gives coding agents access to live production context. Engineers reach it through web, Slack, Teams, MCP and an API. It deploys read-only by default with no agents, schemas or sidecars, and supports bring-your-own-cloud and bring-your-own-model setups. On November 3, 2026 the company announced a $120M Series B at a $1B valuation, co-led by Kleiner Perkins, Redpoint Ventures and Khosla Ventures, for $173M in total funding. Named customers include Nubank, Harvey, DigitalOcean, PepsiCo, Cloudways and American Express. Traversal reports monthly investigations rising from 513 to nearly 73,000 during 2026, and about two-thirds of usage now falls outside root cause analysis. It competes with Resolve AI, Cleric and Datadog Bits AI SRE. Traversal describes self-driving production as its goal; Securities.io says that remains an objective and names the next commercial test as whether customers will let agents change live systems.
A VP of Engineering or head of SRE at a large enterprise with thousands of services and several observability tools, whose incident response is slowed by handoffs between teams.
Shorter incident resolution: DigitalOcean reports a 38% MTTR reduction on Traversal's site, and told Fortune in 2025 that problems were resolved 37% faster over six months.
At a Glance
- Category
- AI Agents & Orchestration
- Pricing
- Contact for pricing
- Target Market
- CTOs, VPs of Engineering, Heads of SRE, Platform Engineering Teams
- Deployment
- Cloud-first, Self-hosted
- Founded
- 2023
- Headquarters
- New York, United States
Key Features
- ✓Production World Model
A continuously updated model of production built from telemetry, code, deployment history and team knowledge, which gives agents context beyond a single dashboard.
- ✓Causal Search Engine
Tests many hypotheses in parallel and follows causal links across systems, so an investigation ends on an evidence-backed root cause.
- ✓Traversal Workers
Long-running agents that start when an incident or alert fires, gather context, identify root causes and help drive remediation.
- ✓Alert Intelligence
Triages incoming alerts so on-call engineers can catch real issues early and spend less time on noise.
- ✓Traversal MCP
Exposes live production intelligence to coding agents, so generated code can account for how services behave in production.
- ✓Read-only default deployment
Runs read-only with no agents, schemas or sidecars, which lowers the risk of a first rollout in a production environment.
Capabilities
Use Cases
- •Major incident root cause analysis
An on-call engineer gets a traced root cause across services, dependencies and recent changes, which cuts the handoffs between owning teams.
- •Release risk review
A team checks a pull request against production behavior, as in Traversal's example of a timeout set below a dependency's P99 latency.
- •Natural-language production support
Engineers ask questions about live production state in Slack or Teams and get answers without writing queries across several observability tools.
- •CMDB validation
Platform teams use the production model to check configuration records against what is running, one of the uses Traversal reports beyond root cause analysis.
Ideal For
Best For
- ✓Root cause analysis across microservice estates with thousands of services
- ✓Alert triage on top of an existing observability stack such as Datadog or Grafana
- ✓Regulated enterprises that need read-only access and bring-your-own-model deployment
- ✓Giving coding agents production context through MCP before they ship changes
Not Ideal For
- ✗Startups and mid-size teams that want a self-serve trial, since access runs through a demo-led enterprise sales process
- ✗Teams looking for a single tool to replace their stack, since Traversal has no log management, metrics collection, tracing, on-call scheduling or status pages
- ✗Organizations with weak instrumentation, since results depend on the telemetry it can read and it does not add what is missing
Deployment
Market Analysis
Pros
- ✓Named production deployments at large, complex estates, including Nubank (5,000+ microservices and over 1 PB of logs per day)
- ✓Read-only default with no agents or sidecars, which keeps the first rollout low risk
- ✓SOC 2 Type 2 and GDPR are listed on its Trust Center, with subprocessors published
Cons
- ✗No public pricing or self-serve path, which makes it hard for smaller teams to evaluate
- ✗No built-in observability or incident management, so it adds to the existing tool bill and does not replace any of it
- ✗Better Stack's comparison says the causal inference is hard to audit, which matters to teams that want step-by-step reasoning
- ✗Most performance figures are vendor-reported or customer quotes selected by the vendor, and revenue is undisclosed beyond a 9x growth claim
Pricing
Enterprise
Contact for pricing
- ✓Alert Intelligence
- ✓Incident RCA
- ✓Production Support
- ✓Self-healing
- ✓Code Resilience
Traversal publishes no list prices and sells through a demo-led enterprise process. One third-party comparison, wetheflywheel.com, describes the model as outcome-driven and not based on data volume, but gives no figures, so budget numbers require a sales call.
Security & Compliance
Sources
This page was written from 9 sources, 7 on domains other than traversal.com.
- 1.traversal.com — traversal.comvendor
- 2.traversal.com — traversal series b 120 million ai srevendor
- 3.trust.traversal.com — trust.traversal.com
- 4.thenextweb.com — traversal 120m series b 1b valuation ai sre
- 5.securities.io — ai sre startup traversal lands 120m series b at 1b valuation
- 6.fortune.com — traversal emerges from stealth with 48 million from sequoia
- 7.betterstack.com — traversal alternatives
- 8.metoro.io — traversal ai alternatives
- 9.wetheflywheel.com — cleric vs resolve ai vs traversal
Stay Ahead of the Curve
Weekly enterprise AI insights for technology leaders. No spam, no vendor pitches—unsubscribe anytime.
SubscribeRelated Products
Tasklet
Cloud AI agents that run recurring business work from plain-English instructions, with no flowchart to build
Voiceflow
AI agent platform for customer support teams, across chat, phone, SMS and WhatsApp
Strawberry
AI browser for sales, recruiting and support teams, with shareable agent 'companions' that work inside your logged-in tabs
Temporal
Open-source durable execution platform that keeps long-running AI agents and workflows alive through crashes, timeouts and failed tool calls