> ## Documentation Index
> Fetch the complete documentation index at: https://docs.sureops.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# What is sureops

> Trust-first AI SRE — from Supervised to Self-Driving incident response

sureops is an AI-powered incident management platform that works alongside your on-call team. It monitors your observability stack around the clock, catches anomalies as they happen, and drives incidents through a structured six-stage pipeline — from first alert to root cause analysis and closure.

The defining characteristic of sureops is its **trust-first model**. It doesn't ask you to hand over the keys on day one. When you first connect your environment, sureops starts in **Supervised mode**: every action it proposes requires your approval before anything executes. You watch it work, calibrate it to your environment, and correct it when it gets something wrong. Over time, as confidence builds on both sides, you can configure it toward **Self-Driving mode**, where the AI drives incidents autonomously — executing runbooks, rolling back deployments, validating recovery — while you remain in the loop via real-time notifications and can intervene at any point.

<CardGroup cols={2}>
  <Card title="Supervised mode" icon="eye">
    sureops proposes every action. You approve before anything executes. Full
    visibility, full control — the AI earns trust incrementally.
  </Card>

  <Card title="Self-Driving mode" icon="bolt">
    sureops drives the full incident lifecycle autonomously. You receive
    real-time updates and retain one-click override at every stage.
  </Card>
</CardGroup>

<Note>
  The progression from Supervised to Self-Driving is **configurable, not
  automatic**. You decide when and how much autonomy to grant — per environment,
  per severity level, and per action type. See [Control modes](/next/getting-started/core-concepts-control-modes) for the full model.
</Note>

sureops integrates with the tools your team already uses — Grafana, Prometheus, Loki, Tempo, Kubernetes, GitHub, Slack, and ArgoCD in private beta — via a vendor-neutral MCP transport layer. Agents read telemetry on-demand through these integrations during live incidents and discard the raw data afterward; your observability data never leaves your cluster.

***

Next: [Core concepts: the 6-stage lifecycle](/next/getting-started/core-concepts-lifecycle)