---
title: "Ask for the outcome. Build the whole flow."
canonical: "https://bird.com/ai/agents"
---

# Ask for the outcome. Build the whole flow.

Give your agent a connected workspace, a clear task, and the context to follow it through.

[Get Started](/dashboard/signup)

[Contact Sales](/demo)

## From prompt to delivered.

Your agent can set up channels, manage customers, and confirm delivery. Full read and write access through MCP and CLI.

**MCP**: `https://mcp.bird.com`

```json
{
  "name": "voice_legs_list"
}
```

Examples: [MCP](/ai/agents.mcp.md) · [CLI](/ai/agents.cli.md) · [TypeScript](/ai/agents.ts.md) · [Python](/ai/agents.py.md) · [Go](/ai/agents.go.md) · [PHP](/ai/agents.php.md) · [HTTP](/ai/agents.http.md)

## A useful first run.

Make the workspace, task, and expected result explicit.

-

### Connect your agent

Choose a supported MCP client or install the CLI. Reuse an existing Bird connection if you already have one.

-

### Confirm the workspace

Read your identity and workspace. Check which sender or resource the task should use.

-

### Ask for a complete outcome

Include the recipient, content, and result you want checked. Let the agent show the operation and its response.

[Set up your agent](/docs/ai/set-up-your-agent)

## Make the request worth following.

Start with one concrete outcome, then connect the next step.

Understand a delivery issue

```
Check my recent email failures. Group them by reason, inspect the affected sending domain, and tell me what needs fixing before making changes.
```

Prepare a new channel

```
Help me set up SMS for order updates. Check my workspace and available senders, explain what is missing, then prepare a test for a number I provide.
```

See what is happening now

```
List the Voice calls that are active right now. Separate connected calls from calls that are still ringing, and show their IDs.
```

## Keep building. Everything connects.

Take the next step with the tools and guides for your project.

- [MCP server](/ai/mcp) — Connect through your preferred AI tool.

- [CLI](/ai/cli) — Run Bird from scripts and your terminal.

- [Skills](/ai/skills) — Give your agent the product context.



## Related resources

- [Setting up your coding agent](/learn/basics/setting-up-your-coding-agent) (video)
- [What is an MCP server, and how does an agent use one to send messages?](/explained/platform/what-is-an-mcp-server-and-how-does-an-agent-send-messages) (answer)
- [Build with AI agents](/learn/paths/agents) (course)

[Get an implementation brief](/learn/workspace?topic=agents)