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

# Legal agents

> Run autonomous background agents that deliver results to the Autopilot inbox.

<Warning>
  Legal Agents require a Vertex or Zenith plan and the legalAgents entitlement. This feature is not
  available on Origin, Learn, or Edge.
</Warning>

Legal Agents are autonomous agents that run in the background and deliver results to the **Autopilot inbox** on the home dashboard. They are not the same as Custo. Agents run without requiring you to remain in the interface.

## The four specialist agents

| Agent               | What it does                                                              |
| ------------------- | ------------------------------------------------------------------------- |
| Document Review     | Clause risk scoring, playbook deviation analysis, redline suggestions     |
| Regulatory Response | Obligation monitoring, regulator enquiry drafts, policy impact assessment |
| Matter Close        | Matter wind-down checklists, outstanding tasks, archive readiness         |
| Client Onboarding   | Client intake orchestration, KYC collection, onboarding pack generation   |

## Human approval gates

Agents do not act unilaterally on sensitive actions. Before proceeding, the agent shows a confirmation prompt explaining what it is about to do. Review and approve or cancel.

## The Autopilot inbox

Completed agent deliverables appear in the Autopilot inbox on the home dashboard. Open each item to review the output, accept recommendations, or dismiss.

## Start an agent run

<Steps>
  <Step title="Open Agents">Go to **Agents** in the sidebar or Launchpad.</Step>
  <Step title="Select an agent">Choose the agent type that matches your task.</Step>
  <Step title="Provide context">Link a matter or document and enter your instruction.</Step>
  <Step title="Launch">Click **Start run**. The agent works in the background.</Step>
</Steps>

## Run history

Open **Agents**, then **Run history**, to see past agent runs, their status, and outputs. Failed runs show the error message and can be retried.

<CardGroup cols={2}>
  <Card title="Custo" icon="message" href="/ai-and-generation/copilot">
    Interactive matter-aware AI assistant.
  </Card>

  <Card title="Plans and limits" icon="credit-card" href="/billing/plans-and-limits">
    Vertex and Zenith entitlements.
  </Card>
</CardGroup>