> 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/serve/text-service.md).

# Text Customer Service: From Setup to Launch

Six steps to take text support from zero to go-live: create an Agent, prepare a knowledge base, connect a conversation platform, adjust platform settings, set up human handoff, and finish with a go-li

Text support lets an Agent reply to customers directly on conversation platforms such as your website, LINE, FB Messenger, and Email, and hand off to a human support agent when it can't answer or the customer asks for one. This page lists six steps in go-live order. The details of each step are on the corresponding MaiAgent Build feature page.

{% hint style="info" %}
Steps 1–5 are done by the builder (IT or Agent administrator), who needs permissions for the relevant Agent, knowledge base, and conversation platform. See [Role Permission Management](https://docs.maiagent.ai/maiagent-user-guide/en/org/roles/role-permission.md). We recommend involving a support supervisor in step 6.
{% endhint %}

## Go-Live Steps <a href="#setup-steps" id="setup-steps"></a>

{% stepper %}
{% step %}

### Create an Agent <a href="#create-agent" id="create-agent"></a>

[Create an Agent](https://docs.maiagent.ai/maiagent-user-guide/en/build/agent/setup.md), and use the [Role Instructions](https://docs.maiagent.ai/maiagent-user-guide/en/build/agent/system-prompt.md) to state clearly which company it represents, the scope it can answer, and its tone. You can also apply an official template from the [Agent Template Library](https://docs.maiagent.ai/maiagent-user-guide/en/build/agent/agent-templates.md) and then modify it. To have the Agent handle a certain type of task by a fixed process (for example, looking up orders or opening tickets), first [create a skill](https://docs.maiagent.ai/maiagent-user-guide/en/build/skills/skills-management.md) and then [bind it to the Agent](https://docs.maiagent.ai/maiagent-user-guide/en/build/skills/skills-management.md#bind-skill-to-ai-agent). Skills travel with the Agent, so every conversation platform you connect later will use them.
{% endstep %}

{% step %}

### Prepare a knowledge base <a href="#prepare-knowledge-base" id="prepare-knowledge-base"></a>

[Create a knowledge base](https://docs.maiagent.ai/maiagent-user-guide/en/build/knowledge-base/km-basic-settings.md) and upload product documents. For questions with fixed answers that need precise replies, build them as [FAQs](https://docs.maiagent.ai/maiagent-user-guide/en/build/knowledge-base/faq.md). Website content can be imported with a [crawler](https://docs.maiagent.ai/maiagent-user-guide/en/build/knowledge-base/scrape-website.md). Use [Search Test](https://docs.maiagent.ai/maiagent-user-guide/en/build/knowledge-base/test-search-result.md) to confirm that the questions customers ask most often find the right passages, and then attach the knowledge base to the Agent.
{% endstep %}

{% step %}

### Connect a conversation platform <a href="#connect-platform" id="connect-platform"></a>

In [Publish to Conversation Platforms](https://docs.maiagent.ai/maiagent-user-guide/en/build/choose-channel.md), compare the credentials each platform requires. Connect the one your customers use most first, for example [Website](https://docs.maiagent.ai/maiagent-user-guide/en/build/choose-channel/web-chat/website.md) or [LINE](https://docs.maiagent.ai/maiagent-user-guide/en/build/choose-channel/line.md), and select this Agent to reply. One Agent can be connected to multiple platforms at the same time: create one conversation platform for each platform and select the same Agent, and the role instructions, knowledge bases, and skills are shared, so there is nothing to redo. After connecting each platform, send a real message on that platform to confirm you get a reply before connecting the next one.
{% endstep %}

{% step %}

### Adjust platform settings <a href="#adjust-platform-settings" id="adjust-platform-settings"></a>

Each platform has its own settings: the greeting and conversation starter questions are in [Interaction Settings](https://docs.maiagent.ai/maiagent-user-guide/en/build/conversation-settings/platform-settings/interaction-settings.md), switches such as file upload and cited sources are in [Function Settings](https://docs.maiagent.ai/maiagent-user-guide/en/build/conversation-settings/platform-settings/function-settings.md), and the hours when AI replies are in [Auto-Reply Settings](https://docs.maiagent.ai/maiagent-user-guide/en/build/choose-channel/web-chat/website.md#auto-reply-settings). For what each tab controls, see [Conversation Platform Settings Overview](https://docs.maiagent.ai/maiagent-user-guide/en/build/conversation-settings/platform-settings.md).
{% endstep %}

{% step %}

### Set up human handoff <a href="#set-up-handoff" id="set-up-handoff"></a>

In [Human Handoff Settings](https://docs.maiagent.ai/maiagent-user-guide/en/build/conversation-settings/human-handoff.md), turn on handoff and fill in the handoff conditions, service hours, and assignment method. Then use [Human Handoff Notifications](https://docs.maiagent.ai/maiagent-user-guide/en/build/conversation-settings/human-handoff/handoff-notification.md) so support agents receive alerts. If you want the AI to draft replies first and have a support agent send every reply, use [Support Copilot](https://docs.maiagent.ai/maiagent-user-guide/en/serve/copilot.md) instead.
{% endstep %}

{% step %}

### Go-live check and quality monitoring <a href="#launch-and-monitor" id="launch-and-monitor"></a>

Before going live, test with real customer questions, or create a [test dataset](https://docs.maiagent.ai/maiagent-user-guide/en/build/agent-ops/test-datasets.md) and run an [automatic evaluation](https://docs.maiagent.ai/maiagent-user-guide/en/build/agent-ops/evaluations.md). Also trigger a human handoff once to confirm that support agents receive it. After going live, check regularly following [AI Support Quality Management](https://docs.maiagent.ai/maiagent-user-guide/en/serve/ai-quality-guide.md), use [Conversation Platform Analytics](https://docs.maiagent.ai/maiagent-user-guide/en/serve/daily-operations/reply/inbox-analysis.md) to find common questions, and [add conversations the Agent answered poorly to the FAQ](https://docs.maiagent.ai/maiagent-user-guide/en/build/agent-ops/add-to-faq.md).
{% endstep %}
{% endstepper %}

## After Go-Live <a href="#after-launch" id="after-launch"></a>

* For the daily work of support agents, such as viewing all conversations, taking over from and handing back to the AI, and looking up contacts, see [Daily Operations for Support Agents](https://docs.maiagent.ai/maiagent-user-guide/en/serve/daily-operations.md).
* To connect additional conversation platforms, repeat steps 3–5. One Agent can serve multiple platforms at the same time.
* To push conversation events to your own systems, see [Webhook](https://docs.maiagent.ai/maiagent-user-guide/en/build/conversation-settings/webhook.md).

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

* [Use Case: Text Support](https://docs.maiagent.ai/maiagent-user-guide/en/serve/text.md): regulation lookup assistant, product inquiry assistant


---

# 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/serve/text-service.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.
