> For the complete documentation index, see [llms.txt](https://en.help.firstline.cc/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://en.help.firstline.cc/feature/ai-agent.md).

# AI Agent

Learn what FIRST LINE's Fancy AI agent can do, and how to configure reply instructions, knowledge, and handoff flows.

{% hint style="info" %}
Fancy AI agent is expected to be available in **Q4 2026** for use. For the actual launch time and feature details, please refer to FIRST LINE's official announcement.
{% endhint %}

## Meet Fancy

{% columns %}
{% column width="16.666666666666664%" %} <img src="https://842546780-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-MNRu7dk70ei7JV9HlW9%2Fuploads%2FHDt6uuM3vZt7pCqCG9wq%2Fai_assistant_anime.png?alt=media&#x26;token=194d5050-36e4-4870-a726-df278c63f686" alt="" data-size="original">
{% endcolumn %}

{% column width="83.33333333333334%" %}
**Fancy is an AI agent from FIRST LINE powered by advanced large language models, designed to understand customer needs, carry on multi-turn conversations, and help handle everyday customer service tasks.** Based on the customer's question, Fancy can determine the appropriate way to respond, retrieve available knowledge content, and, when needed, hand the conversation over to a human agent or a designated conversation script, allowing automated replies to work alongside existing service workflows.
{% endcolumn %}
{% endcolumns %}

Fancy is FIRST LINE's **AI agent**and can understand customer questions, continue responding based on the conversation context, and provide assistance according to the way the business has configured it. When a request requires a fixed procedure, Fancy can hand the conversation over to a conversation script; when human handling is needed, it can also transfer to a human agent.

For example, if a customer first asks, 'Can I return this product?' and then follows up with, 'Can I do that after opening it?' Fancy can continue the previous topic and answer based on the information provided by the business, without the customer having to explain everything again each time.

{% hint style="info" %}
Fancy's responses are affected by the information and settings provided by the business, and may still contain incomplete or incorrect content. Important prices, refund terms, and service commitments should be based on the business's actual policies.
{% endhint %}

## What tasks can Fancy help with?

| Feature                                             | How it can be used                                                                                                     |
| --------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------- |
| **Understand questions and continue conversations** | Answer customers' natural-language questions and maintain context across the conversation.                             |
| **Look up company knowledge**                       | Reference relevant information from the knowledge base, products, and AI training resources.                           |
| **Follow service instructions**                     | Set the brand tone, response style, and principles for handling special situations.                                    |
| **Hand off to fixed workflows**                     | When customers need to inquire about orders or submit requests, transfer them to a prebuilt conversation script.       |
| **Transfer to human support**                       | According to customer needs or business settings, hand conversations that require human assistance to a support agent. |

## Get started with Fancy

{% stepper %}
{% step %}

### Create an AI agent

Log in to FIRST LINE and open the **AI agent**management page, add Fancy, and enter its name and description.
{% endstep %}

{% step %}

### Set basic information

Choose a conversation avatar, and set the idle timeout, human-handoff keywords, the number of consecutive unanswered turns, and the maximum number of conversation turns as needed.
{% endstep %}

{% step %}

### Write response instructions

Go to **Instructions**and set the principles that all conversations should follow, such as tone of voice, what information needs to be confirmed, and what cannot be promised on your own. For special cases, you can add scenario-specific instructions separately.
{% endstep %}

{% step %}

### Select the knowledge scope

Go to **Knowledge**and choose the knowledge sources Fancy should reference. You can select the knowledge base, products, and AI training resources as needed, and use the filters on the screen to narrow the scope.
{% endstep %}

{% step %}

### Set the situations that need to be handed off

Go to **Routing**and set which customer requests should be transferred to a human or handled by a designated conversation script. For example, when a customer wants to formally submit a return request, switch to the return request script.
{% endstep %}

{% step %}

### Test chat and activate

First use**test chat**to check common questions, follow-up questions, and handoff scenarios. After confirming the responses meet your service needs, activate Fancy and select this AI agent in the messaging channels where you want to use it.
{% endstep %}
{% endstepper %}

## Common settings explained

### Basic settings and human handoff

| Setting                      | Description                                                                                                                                                                                                                       |
| ---------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Name, description            | Helps distinguish the purpose of different AI agents.                                                                                                                                                                             |
| Conversation avatar          | Determines the image Fancy displays in conversations.                                                                                                                                                                             |
| Enable / disable             | Controls whether the agent provides service.                                                                                                                                                                                      |
| Idle timeout                 | When the customer has no interaction for a long time, end the conversation according to the settings.                                                                                                                             |
| Human-handoff keywords       | When customers mention specified words, hand them over to human support according to the rules.                                                                                                                                   |
| Consecutive unanswered count | When Fancy fails to find a suitable answer several times in a row and reaches the set count, it will automatically transfer to human support. It is recommended to update knowledge content regularly to reduce unanswered cases. |
| Maximum conversation turns   | After the conversation reaches the set number of turns, transfer it to a human according to the rules.                                                                                                                            |

### Instructions: Decide how Fancy responds

**Default instructions**Used to set Fancy's usual response principles;**Scenario-specific instructions**can supplement how to respond when specific issues arise. Fancy will determine how to apply these instructions based on the conversation content. After setting them, it is recommended to confirm the effect through 'test chat' first. Requests that require fixed steps are better handled with conversation scripts.

For example:

> You are the online customer service for this store. Please respond to customers in Traditional Chinese, clearly and politely. For return and exchange issues, prioritize the provided service rules; if the information is insufficient, confirm the necessary details first and do not promise refund results on your own.

It is recommended to keep product information, prices, and service policies in**Knowledge**for maintenance, so the instructions can focus on explaining 'how to respond.' For detailed examples, see [How do you write AI agent instructions correctly?](https://help.firstline.cc/feature/ai-agent/ru-he-zheng-que-zhuan-xie-ai-dai-li-zhi-shi).

### Knowledge: Provide the information needed for responses

In **Knowledge** the tab, select the content Fancy can reference, such as FAQs, product data, or AI training resources. If you only want to use specific categories or content, you can filter using the conditions provided on the screen.

**Please update knowledge content regularly.** If the data is incomplete, Fancy may not be able to give a definite answer.

### Routing: Decide when to hand off

In **Routing** Set situations and handling methods in the tab, for example:

* When the customer clearly submits a return request,**switch to a conversation script**, and collect information according to the fixed steps.
* When the customer asks for a human or encounters an issue that requires human handling,**Transfer to human support**.

Fancy will determine whether the configured conditions are met based on the conversation content. It is recommended to use clear and specific wording, and confirm the result through test chat.

## When should you use it together with a conversation script?

Fancy is suitable for answering more open-ended questions; conversation scripts are suitable for situations that require fixed menus, data collection, or clear handling steps. The two can be used together:

**Customer asks a question → Fancy understands the need → the conversation script executes the fixed workflow → service ends or continues according to the settings.**

Please note:**After a conversation script completes normally, Fancy will take back the conversation, but it will wait until the customer sends the next message before continuing to reply.** If you want Fancy to**immediately**continue responding, you can use the **"Return to AI Agent"** feature in the conversation script; if the script explicitly ends the conversation or transfers it to a human, Fancy will not take over. Please test the full flow before going live.

Further reading:[Learn about handing off an AI agent to a conversation script](https://help.firstline.cc/feature/ai-agent/liao-jie-ai-dai-li-zhuan-jiao-jiao-tan-jiao-ben),[How can Fancy work with conversation scripts?](https://help.firstline.cc/feature/ai-agent/start-ai-agent-from-scenario).

## Recommendations before use

* First confirm which issues Fancy should handle and which matters should be handed to human support.
* Organize the latest product information and service policies into the knowledge sources.
* Keep the instructions clear and concise, and avoid conflicting requirements.
* Before going live, test scenarios such as general questions, insufficient information, follow-up questions, handoffs, and conversation ending.
* Refunds, fees, contracts, and other important matters should still be handled according to the business-approved service process.
