# AI and automation

Use Brain to ask questions about your telemetry, explain problems and propose changes that a person approves, with the model provider you choose.

Brain answers questions about your data and drafts changes for a person to approve. Workflows and agents automate the work around it, and Auto-Fix turns production errors into pull requests you review. Everything runs on a model provider your admins connect, and one switch turns it all off.

- [Brain](https://tracix.io/docs/ai/brain): Ask about logs, metrics, traces and devices, and confirm the changes Brain proposes

- [Model providers and the AI switch](https://tracix.io/docs/ai/model-providers): Connect Anthropic, OpenAI or a compatible endpoint and control where AI runs

- [Workflows, agents and Auto-Fix](https://tracix.io/docs/ai/workflows): Build workflows with approval steps and review fixes as pull requests

## How Tracix keeps people in charge

* **Your model, your account.** AI features call the provider and model your admins choose. Tracix does not provide a model of its own.
* **Reads use your permissions.** Brain sees only what the person asking can see.
* **Changes wait for a person.** Brain proposes each change in a card you confirm. In workflows, network configuration changes and MCP servers with write tools run only behind an approval step. Auto-Fix opens a pull request only after an approver says yes, and it never merges.
* **One switch stops it all.** An admin can turn off every call to a model or embedding provider in the workspace from **AI Settings**.

## Where to find it in the console

Everything on these pages lives under **AI & Automation** in the console sidebar: **Brain**, **Workflows & Agents**, **Playbook library**, **Auto-Fix**, **Built-in agents** and **AI Settings**.