> ## Documentation Index
> Fetch the complete documentation index at: https://docs.rootly.com/llms.txt
> Use this file to discover all available pages before exploring further.
# AI Settings
> Govern Rootly AI from AI & Agents: the team-wide opt-in, per-feature toggles, AI SRE access, and what each connector can read or change.
## Overview
**AI & Agents** is the customer-facing hub for Rootly AI configuration. Use it to manage self-service feature toggles, instructions, AI SRE rules and Memory, and connector access that is available for your account. Rootly manages AI SRE enablement and some preview controls separately; those controls are called out below.
***
## Where AI Settings Live
Rootly is consolidating the AI configuration surfaces behind a single **AI SRE** item in the sidebar. The settings themselves are unchanged; only where you open them differs.
| Navigation | What you see |
| - | - |
| Consolidated **AI SRE** item | Two tabs: **Investigations** for the investigations AI SRE has run, and **Atlas** for the controls behind them — **AI SRE** and **Conversations** as entry points, **Connectors**, **Private agents**, **Knowledge graph** and **Memory** as context and tools, **Instructions** and **Investigation rules** as investigation settings, and **Global** for organization-wide AI settings. |
| **AI & Agents** | The AI configuration hub described on this page, with **Investigations** as its own sidebar item where that page is enabled. |
The consolidated **AI SRE** navigation is rolling out and may not be enabled for your team yet. This page uses the **AI SRE → Atlas** paths; if your sidebar still shows **AI & Agents**, open the same settings there.
***
## Global
These settings govern Rootly AI's self-service feature surfaces. They do not enable or disable AI SRE, which has a separate Rootly-managed gate, and they do not disconnect configured connectors.
| Setting | What it controls |
| - | - |
| **Opt in to Rootly AI capabilities** | The team-wide switch for self-service Rootly AI features. Turning it off disables those feature surfaces for the current team regardless of their individual toggles. AI SRE doesn't check this setting. |
| **Instructions** | Team-wide guidance inherited by supported Rootly AI features. Configure it under **AI SRE → Atlas → Global → Instructions** when instructions are enabled for your account. More specific [AI SRE instructions](/ai/ai-sre/instructions) take precedence for investigations. |
***
## Features
Rootly AI's core capabilities are grouped by surface area, so your team decides exactly where Rootly AI shows up.
### Incidents
| Toggle | What it controls | Learn more |
| - | - | - |
| **Incident summarization** | AI-generated summaries that bring your team up to speed on the incident. | [AI Summaries](/ai/ai-summaries) |
### Slack
| Toggle | What it controls | Learn more |
| - | - | - |
| **Rootly Agent in Slack** | The Rootly Agent that responders talk to without leaving Slack. The [Chat API](/api-reference/overview#chat-api) also requires this toggle, in addition to its separate team-level enablement. | [Rootly AI in Slack](/ai/rootly-in-slack/overview) |
| **Related incidents** | Enables Slack recommendations and AI SRE's on-demand similar-incidents search tool. AI SRE can still retrieve a small set of similar incidents while preparing hypotheses when this toggle is off. | [Related Incidents](/ai/related-incidents) |
### Retros
| Toggle | What it controls | Learn more |
| - | - | - |
| **AI in Retrospectives** | AI help drafting retrospectives, right inside your templates. | [AI in Retrospectives](/ai/ai-in-retrospectives/overview) |
### Web
| Toggle | What it controls | Learn more |
| - | - | - |
| **Rootly Agent in Web** | The ask-anything AI chat in the web app. | [Rootly AI in Web](/ai/ask-rootly-ai) |
### Mobile
| Toggle | What it controls | Learn more |
| - | - | - |
| **Rootly Agent in Mobile** | The AI chat in the mobile app. | [Rootly AI on Mobile](/ai/rootly-ai-on-mobile) |
***
## AI SRE
Rootly AI SRE runs an investigation on an alert or incident. It tests hypotheses against evidence from your connected tools and reports the most likely root cause when the evidence supports one. Your Rootly account team enables AI SRE for the selected team. Until it's enabled, **AI & Agents → AI SRE** shows a waitlist page. AI SRE is gated separately from the **Opt in to Rootly AI capabilities** setting under **Global**, so that setting doesn't turn AI SRE on or off. See [Getting Started](/ai/ai-sre/getting-started) for availability and first steps.
**AI SRE → Atlas** groups the AI layer into entry points (an **AI SRE** card showing **Enabled** or **Off**, and **Conversations**), the context and tools behind investigations (**Connectors**, **Private agents**), and organization-wide settings (**Global**). Once AI SRE is enabled, these investigation cards appear alongside them:
| Card | What it controls | Learn more |
| - | - | - |
| **Knowledge graph** | A view-only map of the services, infrastructure, and relationships AI SRE uses during investigations. It's part of Atlas, the context layer Rootly AI works from. | [Knowledge Graph](/ai/atlas/knowledge-graph) |
| **Instructions** | Team-wide guidance AI SRE follows in every investigation. | [Instructions](/ai/ai-sre/instructions) |
| **Investigation rules** | Rules that decide when AI SRE investigates an alert and, optionally, the instructions it follows. | [Investigation Rules](/ai/ai-sre/investigation-rules) |
| **Memory** | Durable notes for the selected team that AI SRE can recall during investigations. It's also part of Atlas. | [Memory](/ai/atlas/memory) |
**Instructions**, **Knowledge graph**, and **Memory** may not be enabled for every team. Your Rootly account team turns them on. See the [Atlas overview](/ai/atlas/overview) for how these parts fit with AI connectors and private agents.
Members with an Incident Response seat can open **AI SRE → Atlas** to view these cards (**AI & Agents → AI SRE** if your sidebar doesn't have an **AI SRE** item). Management authority differs by surface and role, and admins without a seat can still reach some surfaces through direct paths; see the [permissions matrix](/managing-users/user-permissions#ai-agents-and-ai-sre). See the [Rootly AI SRE overview](/ai/ai-sre/overview) for the complete investigation flow.
***
## Connectors
You choose the providers Rootly AI can query during investigations and which supported connector actions it can use: observability platforms, code hosts, docs, and your own MCP servers. Most built-in connectors limit calls to a Rootly-reviewed list of tools. Atlassian, Notion, Linear, Braintrust, Honeycomb, AWS, and the Cloudflare connectors pass through the provider's full tool catalog, so the scopes, roles, and permissions you grant in the provider set the boundary. The Custom MCP connector limits calls to the tools you allowlist. Built-in connector tools can accept provider-defined commands or queries that change data, so expose only what's safe for unattended use. For who can view and manage connectors, see the [permissions matrix](/managing-users/user-permissions#ai-agents-and-ai-sre). See [AI Connectors](/ai/connectors/overview) for the complete setup guidance.
***
## Related Pages
What Rootly AI is and everything it can do.
What Rootly AI sees, retention, and model training controls.
Every data source Rootly AI can investigate with.
Investigate alerts and incidents with AI SRE.
The context layer AI SRE and Rootly AI draw on, and where each part lives.
This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.