Create Your First Wize AI Agent

In this guide, we will show you how to create your first Wize AI Agent workflow in REVE Chat.

Today, we build a basic AI Agent workflow that receives a customer message from your website chat, processes the message using an AI Agent, and sends the AI-generated response back to the customer.

This is the basic foundation of every Wize AI Agent workflow. Once you understand this setup, you can later add more advanced tools such as Knowledge Base, Google Sheets, conditions, human handover, Shopify store setup, or multiple AI Agents based on your business needs.

What You Will Learn

By following this guide, you will learn how to:

• Create a new Wize AI Agent workflow
• Add and configure a chat trigger
• Add an AI Agent node
• Connect the customer’s message to the AI Agent
• Write a basic AI Agent prompt
• Send the AI Agent response back to the customer
• Test, save, and activate the workflow

What We Will Build

In this guide, we will build a simple AI Agent workflow with three main parts:

On chat message → AI Agent → Response To Chat

This means:

• The workflow will start when a customer sends a message.
• The AI Agent will receive and process the customer’s message.
• The final AI-generated response will be sent back to the customer through chat.

Before You Begin

Before we create the workflow, make sure you have:

• An active REVE Chat account
• Access to the REVE Chat dashboard
• A website chat widget or communication channel configured
• A clear idea of what your AI Agent should help customers with
• Basic instructions for how the AI Agent should respond

For this first workflow, you do not need to connect a Knowledge Base or Google Sheet. We will first create a basic AI Agent workflow, and later you can add business data, tools, conditions, and human handover.

Step 1: Go to AI Agent workflows

First, go to your REVE Chat dashboard.

From the left-side menu, navigate to: Automation > AI Agent > Create Workflow

Here, you will see the AI Workflows page. This page displays all your existing AI Agent workflows, including the workflow name, creator, creation date, last updated date, and current status.

👉 To create a new workflow, click the ”+ Create Workflow” button.

Step 2: Add workflow details

Next, you need to add the basic details for your workflow.

In the Create new workflow pop-up, fill in the following fields:

After entering the details, click Continue.

Once you continue, the visual workflow builder will open. This is where we will build the AI Agent workflow using drag-and-drop nodes.

Step 3: Setting Up a Trigger

Every workflow needs a trigger. A trigger tells the workflow when it should start.

In this step, we will add a trigger so the AI Agent can start working when a customer sends a message.

Click Add Trigger and choose how you want the workflow to start.

Available trigger options include:

For this guide, select On chat message.

This is the best option for your first AI Agent workflow because we are creating an AI Agent that responds to customer messages from the chat widget.

Step 4: Configure the On chat message trigger

Now we will configure the On chat message trigger.

Click the On chat message trigger node to open its configuration panel.

In the Inputs tab, configure the following fields:

After configuring the trigger, close the configuration panel.

At this stage, your workflow can now start when a customer sends a message from the selected chat source.

Step 5: Add the AI Agent node

Now we will add the AI Agent node.

The AI Agent node is the main part of the workflow. It receives the customer’s message, follows your instructions, uses the selected AI model, and generates a response.

To add the AI Agent node:

1. Click the + button after the trigger node.
2. Select AI Agent from the node list.

Once added, the AI Agent node will appear after the On chat message trigger.

Step 6: Configure the AI Agent input

Now we need to tell the AI Agent which message it should process.

To do this, we will connect the customer’s chat message from the trigger to the AI Agent’s Input prompt field.

Click the AI Agent node to open its configuration panel.

In the Input prompt field, insert the incoming customer message from the trigger output.

To add the customer message:

1. In the right-side data panel, expand Nodes.
2. Expand On chat message.
3. Go to body > query > message.
4. Press and hold text.
5. Drag and drop text into the Input prompt field.

The input prompt should be added in this format:

{{local.on_chat_message.body.query.message.text}}

This connects the customer’s chat message to the AI Agent, allowing the AI Agent to process the message as input.

📌 Note: If the input data is not visible, keep the input field blank temporarily and configure the remaining AI Agent settings. Then close the panel and run the workflow once from the On chat message node. After a message passes through the node, the input data should become available in the data panel.

Step 7: Select the AI provider and model

Next, select the AI provider and model that the AI Agent will use to process customer messages.
In the Chat model section, configure the AI model.

Step 8: Add the AI Agent prompt message

In the Prompt Message field, write instructions that define how the AI Agent should behave.

The prompt tells the AI Agent what role it should play, what information sources it should use, and how it should respond to customers.

Example prompt:

“You are a professional e-Commerce Store Support Agent. Your primary responsibility is to help customers use information retrieved from the store. Always keep your response professional, friendly, and concise.’’

After adding the prompt, close the AI Agent configuration panel.

Step 9: Add a Response To Chat node

Now we need to send the AI Agent’s answer back to the customer.

To do this, add a Response To Chat node after the AI Agent node.

Click the ‘’+’’ button after the AI Agent node and select Response To Chat.

Your basic workflow should now look like this:

On chat message → AI Agent → Response To Chat

Step 10: Configure the response message

Now we will configure the Response To Chat node.

Click the Response To Chat node to open its configuration panel.

In the Inputs tab, configure the following fields:

To connect the AI Agent output:

1. In the right-side data panel, expand Nodes.
2. Expand AI Agent.
3. Expand structured_response.
4. Drag and drop output into the Message Response field.

The message response should be added in this format:

{{local.ai_agent.output}}

This sends the AI-generated response back to the customer in the chat.

📌 Important: If the AI Agent output is not added to the Message Response field, the workflow may run successfully, but the customer will not receive the AI response in the chat. 

Step 11: Test the Wize AI Agent workflow

After configuring the workflow, test it before activating it.

Use the Test Flow panel to send a sample customer message.

Step 12: Save and activate the workflow

After testing the workflow:

1. Review all connected nodes.
2. Make sure the trigger, AI Agent, and response node are properly configured.
3. Click Save.
4. Turn the workflow status to Active.

Once activated, the Wize AI Agent can start responding to customer messages based on the configured workflow.

Troubleshooting

Best Practices

• Use a clear workflow name so your team can identify it easily.
• Start with a simple workflow before adding advanced tools.
• Write a specific prompt that explains the AI Agent’s role and response style.
• Do not ask the AI Agent to answer from data sources that are not connected yet.
• Always test the workflow before activating it.
• Check the execution log if the workflow does not behave as expected.
• Keep your first workflow simple: trigger, AI Agent, and response node.

What to Add Next

Now that you have created your first Wize AI Agent workflow, you can make it more powerful by connecting business data and workflow logic.

You can add:

• A Knowledge Base to help the AI Agent answer from FAQs, policies, help articles, and support documents
• Google Sheets to retrieve structured data such as product inventory, pricing, ratings, and stock availability
• Conditions to control whether the AI Agent should respond or hand over the chat
• Human Handover to transfer complex conversations to live agents
• Additional AI Agent tools for advanced multi-agent workflows

Next Step

You have now created your first Wize AI Agent workflow. Next, we will show you how to connect a trained knowledge base so your AI Agent can answer using your business-specific content.

👉 Continue to: How to Add Knowledge Base Information in AI Agent