> For the complete documentation index, see [llms.txt](https://docs.maiagent.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.maiagent.ai/maiagent-user-guide/en/build/tools/agent-ui/quick-reply.md).

# Quick Reply

Create a Quick Reply button row so your Agent can attach clickable options to responses and guide users to their next step

Quick Reply is a type of [Agent UI](https://docs.maiagent.ai/maiagent-user-guide/en/build/tools/agent-ui.md). After an Agent responds, a row of clickable buttons (chips) appears below the message, letting users choose their next step without typing: send a message, open a link, or prefill the input box. The Agent uses the trigger conditions to decide when to display the buttons, with a maximum of 13 buttons.

## Use cases <a href="#use-case" id="use-case"></a>

A telecom customer service team wants its Agent to ask “How can I help you?” when a user starts a conversation without a clear intent, and attach three buttons: “Check bill,” “Change plan,” and “Report an issue.” When the user clicks “Check bill,” the Agent goes directly to the billing flow. Another option is dynamic generation: when a user discusses a product they are interested in, the LLM lists up to 5 related products as buttons based on the conversation.

Compared with the [Quick Menu](https://docs.maiagent.ai/maiagent-user-guide/en/build/tools/quick-menu.md): the Quick Menu **stays in the input bar** and does not change with the conversation. Quick Reply is **attached below a specific response**. The Agent decides when it appears, and its content can be generated dynamically based on the conversation.

## Create a Quick Reply <a href="#create" id="create"></a>

{% stepper %}
{% step %}

### Select the type <a href="#step-pick-type" id="step-pick-type"></a>

Go to “<mark style="color:blue;">AI Features</mark>” → “<mark style="color:blue;">Agent UI</mark>,” click “<mark style="color:blue;">Add Agent UI</mark>,” and select **Quick Reply** (a row of quick reply buttons).

<figure><img src="https://1360999650-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F6v6TNkkOQVfRYfcNirHL%2Fuploads%2Fgit-blob-b3e3e9a5e71c9b2d4cff4aa99615ba6ff00de631%2Fagent-ui-quick-reply-pick-type.png?alt=media" alt="Quick Reply card on the Agent UI type selection page"><figcaption><p>Click Quick Reply under “Select Agent UI Type”</p></figcaption></figure>
{% endstep %}

{% step %}

### Enter basic information <a href="#step-basic-info" id="step-basic-info"></a>

| Field                                               | Description                                                                                                                                                                                                                                               |
| --------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| <mark style="color:blue;">Display Name</mark>       | The name shown in the admin interface, such as “Customer Service Menu”                                                                                                                                                                                    |
| <mark style="color:blue;">Tool Name</mark>          | An English identifier used by the LLM. It must follow the naming rules (English letters and underscores, with no spaces)                                                                                                                                  |
| <mark style="color:blue;">Description</mark>        | Optional. Describe what this set of buttons is for                                                                                                                                                                                                        |
| <mark style="color:blue;">Trigger Conditions</mark> | **Required.** Tell the Agent in natural language when to attach this set of buttons, such as “When a user starts a conversation without a clear intent and needs help choosing” or “Provide category shortcuts when users ask frequently asked questions” |

{% hint style="info" %}
The more specific the trigger conditions, the better the Agent can display the buttons at the right time instead of attaching them to every response. If you attach multiple Quick Reply sets to one Agent, keep their trigger conditions from overlapping.
{% endhint %}
{% endstep %}

{% step %}

### Select a template (optional) <a href="#step-template" id="step-template"></a>

The “<mark style="color:blue;">Select Template</mark>” section lists built-in templates. Click one to preview it on the right. Clicking “<mark style="color:blue;">Apply This Template</mark>” **overwrites** the current button settings with the template content. You can edit the settings afterward.

<figure><img src="https://1360999650-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F6v6TNkkOQVfRYfcNirHL%2Fuploads%2Fgit-blob-9b43f0f93949382cc0e9ce547df260c1cc31e054%2Fagent-ui-quick-reply-editor.png?alt=media" alt="Quick Reply editor and preview after applying the Customer Service Menu template"><figcaption><p>After you apply the “Customer Service Menu” template, the preview on the right simulates the Web Chat button row</p></figcaption></figure>
{% endstep %}

{% step %}

### Configure buttons <a href="#step-buttons" id="step-buttons"></a>

First, select <mark style="color:blue;">Static List</mark> or <mark style="color:blue;">Dynamic Generation</mark> at the top of “<mark style="color:blue;">Button Settings</mark>” (see the differences in the next section), then configure each button. The “<mark style="color:blue;">Preview</mark>” on the right simulates how the buttons appear in Web Chat in real time. Click the buttons to try them, then click “<mark style="color:blue;">Show Buttons Again</mark>” to start over.
{% endstep %}

{% step %}

### Save <a href="#step-save" id="step-save"></a>

After checking the preview, click “<mark style="color:blue;">Save</mark>” in the upper-right corner. Before saving, the system checks that button text is not empty, fixed text is no longer than 20 characters, AI fields have prompts, links start with `https://`, and the quantity range is between 1 and 13.
{% endstep %}
{% endstepper %}

## Static List and Dynamic Generation <a href="#static-vs-dynamic" id="static-vs-dynamic"></a>

| Mode                                                | How buttons are determined                                                                                                 | Best suited for                                                       |
| --------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------- |
| <mark style="color:blue;">Static List</mark>        | Configure a fixed set of 1–13 buttons. For each button, choose “Fixed” or “AI” separately for its text and action fields   | Fixed service categories and FAQ shortcuts                            |
| <mark style="color:blue;">Dynamic Generation</mark> | Provide a single “button pattern” and a quantity range. The LLM generates N buttons in real time based on the conversation | Product recommendations and options based on topics the user mentions |

### Static List <a href="#static-list" id="static-list"></a>

Click “<mark style="color:blue;">Add Button</mark>” to add a button (up to 13). Expand each button to configure the fields described in the next section. Click the trash icon to the right of its title to delete it. Each button title shows its action type label and button text for quick identification.

### Dynamic Generation <a href="#dynamic-generation" id="dynamic-generation"></a>

| Field                                                                                               | Description                                                                                                                                                                                             |
| --------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| <mark style="color:blue;">Minimum Quantity</mark>／<mark style="color:blue;">Maximum Quantity</mark> | The number of buttons the LLM can generate. Must satisfy 1 ≤ minimum ≤ maximum ≤ 13                                                                                                                     |
| <mark style="color:blue;">Generation Rules</mark>                                                   | Tell the LLM when and how to generate buttons, and how many. For example, “List up to 5 related products based on the interests the user discusses; skip the entire set if there are no relevant items” |
| <mark style="color:blue;">Button Pattern</mark>                                                     | Settings shared by all generated buttons (action type, text, and icon). The text field is usually set to AI generation                                                                                  |

Dynamic Generation mode does not support JSON source editing.

## Settings for each button <a href="#button-fields" id="button-fields"></a>

### Action types <a href="#action-types" id="action-types"></a>

| Action                                         | When clicked                                                                         | Field to fill in                                                                                                |
| ---------------------------------------------- | ------------------------------------------------------------------------------------ | --------------------------------------------------------------------------------------------------------------- |
| <mark style="color:blue;">Send Message</mark>  | Sends a message to the Agent as the user                                             | <mark style="color:blue;">Message to Send</mark>: The text actually sent, which can differ from the button text |
| <mark style="color:blue;">Open Link</mark>     | Opens a URL                                                                          | <mark style="color:blue;">Link URL</mark>: Must start with `https://`                                           |
| <mark style="color:blue;">Prefill Input</mark> | Places text in the input box without sending it. The user can edit it before sending | <mark style="color:blue;">Prefill Text</mark>                                                                   |

### Fixed and AI <a href="#fixed-vs-ai" id="fixed-vs-ai"></a>

“<mark style="color:blue;">Button Text</mark>” and the action fields in the table above each have a <mark style="color:blue;">Fixed</mark>／<mark style="color:blue;">AI</mark> switch on the right:

* **Fixed**: Displays the value you enter. Button text can be up to 20 characters long.
* **AI**: Enter a prompt, and the LLM generates content from it during the conversation. For example, set button text to AI and enter “The name of the product the user just mentioned.”

### Button icon <a href="#button-icon" id="button-icon"></a>

For each button, choose <mark style="color:blue;">None</mark>, <mark style="color:blue;">Image</mark> (enter an `https://` image URL), <mark style="color:blue;">Emoji</mark> (a single emoji), or <mark style="color:blue;">Initials Badge</mark> (1–2 characters).

{% hint style="info" %}
Selecting “<mark style="color:blue;">None</mark>” for “<mark style="color:blue;">Button Icon</mark>” means no custom icon is added. The “<mark style="color:blue;">Preview</mark>” on the right side of the admin interface still shows an action indicator: a paper airplane for <mark style="color:blue;">Send Message</mark>, an external link icon for <mark style="color:blue;">Open Link</mark>, and an edit icon for <mark style="color:blue;">Prefill Input</mark>. Switching templates does not change the icon to one for a different action; to do that, change the “<mark style="color:blue;">Action Type</mark>.” This describes the admin preview. Icon placement and appearance may differ on the visitor side.
{% endhint %}

<figure><img src="https://1360999650-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F6v6TNkkOQVfRYfcNirHL%2Fuploads%2Fgit-blob-2536eab0bf9cbcd865b5264af22ae3b951c17a5d%2Fquick-reply-action-icon-preview-prod.png?alt=media" alt="A paper airplane icon appears to the right of the service inquiry button in the admin Quick Reply preview"><figcaption><p>Admin preview: with Button Icon set to “None” and Action Type set to “Send Message,” the paper airplane action indicator still appears.</p></figcaption></figure>

### Advanced action JSON (LINE only) <a href="#advanced-action-json" id="advanced-action-json"></a>

To use native LINE actions beyond the three basic actions (such as `datetimepicker`, `postback`, `camera`, `cameraRoll`, `location`, `richmenuswitch`, and `clipboard`), click an example tab above to insert one quickly, or paste a LINE action object as JSON. The system passes it to LINE unchanged. These buttons appear gray in the preview and do not appear in Web Chat.

## JSON source <a href="#json-source" id="json-source"></a>

The “<mark style="color:blue;">JSON Source</mark>” section (collapsed by default) corresponds to the LINE `quickReply` JSON for the current static list. Advanced users can edit it directly, then click “<mark style="color:blue;">Apply JSON to Button Settings</mark>.” Applying it overwrites the button settings and supports only the `message`／`uri`／`inputOption` actions. Any buttons beyond the first 13 are discarded.

## Attach to an Agent <a href="#attach-to-agent" id="attach-to-agent"></a>

After saving, attach the Quick Reply to an Agent using **Agent mode** for it to take effect: go to Agent settings → “<mark style="color:blue;">Interface & Experience</mark>” → “<mark style="color:blue;">Agent UI</mark>” → “<mark style="color:blue;">Select Agent UI</mark>,” select the Quick Reply you just created, and save. See [Bind to an Agent](https://docs.maiagent.ai/maiagent-user-guide/en/build/tools/agent-ui.md#bind-to-chatbot) for details.

## FAQ <a href="#faq" id="faq"></a>

<details>

<summary>Which platforms display the buttons?</summary>

Web Chat and LINE. Both support the three basic actions; “Advanced action JSON” works only in LINE.

</details>

<details>

<summary>Why is the Agent not attaching buttons?</summary>

First, check that the Agent's response mode is Agent mode, this Quick Reply set is selected on the Agent UI tab, and the “Trigger Conditions” match the user's actual question. In Dynamic Generation mode, the LLM skips the entire set if it determines there are no suitable items based on the “Generation Rules.”

</details>

<details>

<summary>Do the buttons disappear after a user clicks one?</summary>

Yes. The buttons are attached only to that response. Clicking one hides the entire row (the preview displays “Buttons hidden”). The Agent decides again whether to attach buttons to the next response.

</details>


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.maiagent.ai/maiagent-user-guide/en/build/tools/agent-ui/quick-reply.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
