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

# Manual actions

> Actions that Human Agents perform when the AI Agent requests

Manual actions let the AI Agent request a Human Agent to perform tasks that
can't be automated, either because no API exists or the task requires human
judgment.

**Best for:**

* Systems without API or integration support
* Sensitive tasks requiring human judgment
* Bridging gaps while implementing full integration

## How manual actions work

1. **AI Agent identifies need**: During a conversation, it determines a manual
   action is required
2. **AI Agent collects information**: Gathers required data from the customer
3. **Request sent to Human Agent**: Conversation hands off with a pending action
4. **Human Agent executes**: Performs the action in the external system
5. **Human Agent confirms**: Marks the action done or declines it

## Configure a manual action

![Manual action configuration form](https://mintcdn.com/typewise/zYjp2yUGRaF5N0OL/images/documentation/actions/manual-actions/manual-action-form.webp?fit=max\&auto=format\&n=zYjp2yUGRaF5N0OL\&q=85\&s=85f8cc9ebf81adee1bc0c27c82b7c4e0)

| Field | Description |
| :- | :- |
| **Action name** | Descriptive name (e.g., "Refund Order", "Update Billing Address") |
| **Action description** | When the AI Agent should suggest this action |
| **Data Collection fields** | Data the AI Agent collects before requesting |

### Data Collection fields

Each field has:

* **Name**: Identifier (e.g., `order_id`, `refund_reason`)
* **Description**: What to ask the customer
* **Type**: Text, Number, or Yes/No
* **Required**: Required or Optional

## Human Agent workflow

When the AI Agent requests a manual action, Human Agents see it in their inbox.

### View pending actions

The request appears as a card in the conversation, asking **Have you completed
this action?** It shows:

* **Action**: the action requested (e.g., "Refund Order")
* **Description**: the action's description
* One copyable field per piece of information collected from the customer

![Manual action request card with Decline and Done](https://mintcdn.com/typewise/zYjp2yUGRaF5N0OL/images/documentation/actions/manual-actions/manual-action-request.webp?fit=max\&auto=format\&n=zYjp2yUGRaF5N0OL\&q=85\&s=230acb3d73118694adb54e7d910567cb)

### Done or decline

**To mark it done:**

**1. Review information**

Review the collected information.

**2. Perform action**

Perform the action in your external system.

**3. Confirm**

Click **Done**.

**To decline:**

**1. Click Decline**

Click **Decline**.

**2. Provide reason**

Write a reason (helps the AI Agent respond appropriately). The conversation
returns to AI Agent handling with your feedback.

## Example: Refund order

**Configuration:**

| Field | Value |
| :- | :- |
| Action name | Refund Order |
| Description | Use when a customer requests a refund. Collect order ID and reason first. |

**Data Collection fields:**

| Name | Description | Type |
| :- | :- | :- |
| `order_id` | Order number to refund | Text |
| `refund_reason` | Why the customer wants a refund | Text |
| `refund_amount` | Amount to refund (or "full") | Text |

**Human Agent sees** a card with:

```
Action: Refund Order
Description: Use when a customer requests a refund. Collect order ID and reason first.
order_id: ORD-12345
refund_reason: Item arrived damaged
refund_amount: full
```

The Human Agent processes the refund, then clicks **Done**.

## See also

* [AI actions](https://docs.typewise.app/documentation/actions/ai-actions.md): For automated execution
* [Ask a human agent](https://docs.typewise.app/documentation/actions/ask-a-human-agent.md): For one-off
  requests the AI improvises, with no Actions hub entry behind them


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.