# How to Set Up AI Assistants in Mercuri

### **1. Go to the AI Assistants Section**

1. Log in to your **Mercuri Dashboard**.
2. In the left-side panel, navigate to **AI Assistants**.
3. Click on **“+ Create Agent”** to start setting up a new assistant.

<figure><img src="/files/2vtTG0aSbWqOdzzb5J2u" alt=""><figcaption></figcaption></figure>

***

### **2.** Configure the AI Assistant

<figure><img src="/files/PDdf29fQXiuFhmjebFcP" alt=""><figcaption></figcaption></figure>

Under the **Set Up Your Agent** section, you will find 6 tabs. Configure each tab carefully to set up your AI Assistant accurately.

Let’s go through each tab one by one.

#### <mark style="color:blue;">1. General</mark>

In the **General** tab, enter the basic details of your AI Assistant, such as:

* **Agent Name**
* **AI Model** (pre-selected)
* **System Prompt**

{% hint style="info" %}
The **System Prompt** defines how the AI Assistant should behave and respond to users.

Provide clear instructions about:

* **The assistant’s responsibilities**\
  \&#xNAN;*(e.g., answering customer queries, sharing product information, handling basic support requests)*
* **The communication style**\
  \&#xNAN;*(e.g., professional, concise, friendly, and helpful)*
* **Any limitations or escalation rules**\
  \&#xNAN;*(e.g., directing users to a human agent for complex, sensitive, or account-related issues)*

These instructions help ensure the AI Assistant responds consistently and accurately.

**Example Prompt**

{% code overflow="wrap" %}

```
You are a friendly and knowledgeable assistant for our website. 

You help customers with product information, FAQs, booking questions, and basic support. Keep responses clear, polite, and concise. 

For anything urgent or account-related, direct the user to our support team.
```

{% endcode %}

{% endhint %}

***

#### <mark style="color:blue;">**2. Knowledge**</mark>&#x20;

<figure><img src="/files/5ONDXnKQLtQ1DGCXv7Qt" alt=""><figcaption></figcaption></figure>

Once your AI assistant identity is configured, the next step is to train it with your business information so it can respond accurately to customer queries.

Inside the **Knowledge** section in Mercuri, you can connect different knowledge sources that the AI will use while replying to conversations.

**You can add:**

* Your website URL
* Sitemap URL (`sitemap.xml`)
* Individual pages like:
  * Pricing pages
  * FAQ pages
  * Service pages
  * Product pages
  * Booking or policy pages

<details>

<summary><strong>Why Add a Sitemap?</strong></summary>

Adding your sitemap allows Mercuri to automatically crawl and index your website pages.

This helps the AI assistant:

* Understand your business better
* Learn your services and workflows
* Access FAQs and important information
* Provide more accurate customer responses

Mercuri recommends starting with the sitemap because it gives the AI a complete overview of your website content.

</details>

<details>

<summary><strong>When to Add Specific Pages Instead?</strong></summary>

If you want the AI assistant to focus only on important business pages, you can manually add individual URLs.

This is useful when you want tighter control over AI responses for:

* Pricing information
* Appointment booking details
* Refund or cancellation policies
* Service-specific information
* Support documentation

</details>

<details>

<summary><strong>Benefits of Connecting Knowledge Sources</strong></summary>

Once your knowledge sources are connected, the AI assistant can:

* Answer customer questions instantly
* Give context-aware responses
* Reduce repetitive support queries
* Improve response accuracy
* Deliver a more human-like conversation experience
* Help customers get information faster without manual team involvement

</details>

***

#### <mark style="color:blue;">3. Tools</mark>

<figure><img src="/files/u9gXKt5SMQpdxrccUdn2" alt=""><figcaption></figcaption></figure>

**Tools** section defines what your assistant can do. These are the tools and capabilities that the AI will have access to.

{% tabs %}
{% tab title="Wix Bookings Support" %}
*Helps customers manage their bookings and appointments.*

Use this if your business relies on:

* Appointment scheduling
* Class or session bookings
* Event reservations
* Service-based consultations

Your AI Assistant will help customers with:

* Booking new appointments
* Checking availability
* Rescheduling or cancelling bookings
* Confirming upcoming sessions

This is ideal for salons, coaching centres, clinics, fitness trainers, consultancies, etc.
{% endtab %}

{% tab title="Order Tracking" %}
*Assist customers in tracking their order status and delivery.*

Enable this if you want your AI Assistant to automatically answer:

* “Where is my order?”
* “Has my package been shipped?”
* “When will I receive my delivery?”

The system fetches live shipping and delivery data, reducing workload for support teams.

Best for D2C brands, e-commerce stores, food delivery services, and logistics-based businesses.
{% endtab %}

{% tab title="Product Enquiry" %}
*Answer customer questions about products, features, and availability.*

Useful for businesses that offer multiple products or variants:

* Clothing & accessories
* Electronics
* Food & FMCG
* Health & beauty
* Home decor

Your AI Assistant will help users with:

* Product features
* Stock availability
* Variants and options
* Price and offers
* Recommendations

This automates a major part of pre-purchase conversations.
{% endtab %}

{% tab title="Customer Support" %}
*Provide general customer service and support.*

Enable this for:

* Handling FAQs
* Answering policy-related questions
* Assisting with basic troubleshooting
* Guiding users through workflows

Your AI will act as the first line of support, escalating to human agents when needed.

Best for any business that receives recurring customer service queries.
{% endtab %}
{% endtabs %}

***

***

#### <mark style="color:blue;">5. Channels</mark>

<figure><img src="/files/U7eFUQyBm5ABXKhP24pt" alt=""><figcaption></figcaption></figure>

After setting up your AI assistant and knowledge base, the next step is to connect the communication channels where customers will interact with your business.

Inside the **Channels** section in Mercuri, you can connect platforms like:

* WhatsApp
* SMS

Mercuri allows you to manage these conversations directly through the AI assistant.

#### Connect a WhatsApp / SMS Channel

To connect :

* Select **WhatsApp /** **SMS** from the available channel options
* Choose the phone number you want to connect
* Click + **Connect Channel**

<figure><img src="/files/YtrhM53PBQSphlDdOfJH" alt=""><figcaption></figcaption></figure>

#### Connect Test Number

After clicking on **+ Connect Channel**, Mercuri will prompt you to add a test number where you can test your AI assistant before going live.

**To add a test number:**

* Enter your phone number in the provided field
* Click on **+ Add**

Only the phone numbers added here will be able to chat with your AI assistant while it is in testing mode.

{% hint style="info" %}

#### Why This Is Important

Adding test numbers allows you to:

* Test AI responses safely
* Verify workflows and automations
* Check conversation quality
* Ensure the assistant behaves correctly before customers use it

This helps avoid errors and gives you full confidence before launching the assistant publicly.
{% endhint %}

#### Manage Connected Channels

Mercuri also gives you control over connected channels directly from the dashboard.

You can:

* Enable or disable the channel using the **toggle icon**
* Delete the connected channel using the **bin/delete icon**

This makes it easy to manage testing, pause communication temporarily, or remove channels when needed.

***

#### <mark style="color:blue;">6. Launch</mark>

<figure><img src="/files/NRg8L9x7WaDv8ne9ENpN" alt=""><figcaption></figcaption></figure>

After configuring your AI assistant, adding knowledge sources, and connecting messaging channels, the final step is to train and launch the agent.

Inside the **Launch** section in Mercuri, you can prepare the assistant for live customer interactions.

***

#### Train Your AI Assistant

Before the assistant can respond accurately, Mercuri needs to index and process all the connected knowledge sources.

To start training:

* Click on **Start Training**
* Mercuri will process:
  * Website content
  * Sitemap pages
  * Added URLs
  * Connected knowledge sources

The training process usually takes **1–5 minutes** depending on the amount of content added.

<figure><img src="/files/P2Y4QwCNFQxQin02DQyw" alt=""><figcaption></figcaption></figure>

#### Deployment Modes

Once training starts, Mercuri provides different deployment modes for managing the assistant.

{% tabs %}
{% tab title="Testing Mode" %}
In Testing mode:

* Only added test numbers can interact with the assistant
* You can safely validate responses and workflows
* Customers will not have access yet

This mode is recommended before going live.
{% endtab %}

{% tab title="Live Mode" %}
After successful testing, you can switch the assistant to **Live** mode.

In Live mode:

* Real customers can interact with the AI assistant
* The assistant becomes active on connected channels like WhatsApp or SMS
* Automations and AI replies start functioning publicly

Make sure testing is completed before enabling Live mode.
{% endtab %}

{% tab title="Back to Draft" %}
If you need to make changes later, you can move the assistant back to **Draft** mode.

This allows you to:

* Edit prompts or workflows
* Update knowledge sources
* Modify configurations
* Retrain the assistant after changes

Mercuri makes it easy to continuously improve the assistant even after deployment.
{% endtab %}
{% endtabs %}


---

# Agent Instructions: Querying This Documentation

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

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.mercuri.cx/features/ai-assistants/how-to-set-up-ai-assistants-in-mercuri.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
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.
