n8n Chat Node: Build Conversational Workflows with Human-in-the-Loop

The n8n Chat Node is one of the most powerful additions to n8n in 2026, enabling you to build truly interactive workflows that pause for user input, require human approval before executing critical actions, and create conversational interfaces for everything from customer support to AI agent oversight.

Unlike traditional automation that runs start-to-finish without human interaction, the Chat Node lets you insert conversation points anywhere in your workflow—perfect for scenarios where you need user confirmation, additional context, or manual decision-making before the automation continues.

What Is the n8n Chat Node?

The Chat Node works alongside the Chat Trigger node to create multi-turn conversational workflows. While the Chat Trigger starts the workflow when a user sends a message, the Chat Node allows you to send messages back to the user and optionally pause execution until they respond.

This enables powerful patterns like:

• Human-in-the-loop AI workflows where an AI agent proposes actions but waits for human approval before executing them
• Interactive form collection that asks follow-up questions based on previous answers
• Approval gates for sensitive operations like deleting customer data or processing refunds
• Conversational customer support that escalates to humans when needed
• Internal tools that guide team members through complex processes step-by-step

The Chat Node supports two response types: Free Text (users type any response) and Approval (users click inline approve/disapprove buttons). This flexibility makes it suitable for both open-ended conversations and structured decision points.

Chat Node vs Chat Trigger: Understanding the Difference

If you're new to n8n's chat capabilities, understanding the relationship between Chat Trigger and Chat Node is essential:

Chat Trigger: Starts your workflow when a user sends a message. This is always the entry point for chat-based workflows. You configure whether to use n8n's hosted chat interface or embed your own chat widget.

Chat Node: Sends messages during workflow execution and optionally waits for responses. You can use multiple Chat Nodes in a single workflow to create multi-turn conversations.

Think of it like a phone conversation: the Chat Trigger is the user calling you (initiating contact), and each Chat Node is you responding and potentially asking follow-up questions before continuing the conversation.

Setting Up Your First Chat Workflow

The fastest way to start building with the Chat Node is using n8n's Hosted Chat option, which provides a ready-to-use chat interface without any additional setup.

Step 1: Add a Chat Trigger node

Create a new workflow and add the Chat Trigger node. Under "Chat Options," select "Hosted Chat" (recommended for most users). This gives you a shareable chat URL that you can send to users or embed on your website.

Step 2: Configure your initial message

In the Chat Trigger settings, set your "Initial Messages" to greet users when they open the chat. For example: "Hi! I'm here to help with your order. What's your order number?"

Step 3: Add a Chat Node to respond

Connect a Chat Node after your Chat Trigger. Set the operation to "Send and Wait for Response" and configure your message. The workflow will pause here until the user responds.

Step 4: Process the response and continue

The user's response is available in {{ $json.chatInput }}. You can use this data in subsequent nodes—for example, looking up an order in your database or passing it to an AI agent for processing.

If you're running workflows on n8nautomation.cloud, your chat interface is automatically hosted with 24/7 uptime and no server management required.

Send Message vs Send and Wait for Response

The Chat Node offers two distinct operations that serve different purposes in your conversational workflows:

Send Message

This operation sends a message to the chat and immediately continues workflow execution without waiting for a response. Use this when you want to:

• Provide status updates ("Looking up your order...")
• Send confirmation messages ("Your refund has been processed")
• Deliver information without needing user input
• Show progress in multi-step workflows

Example: After a user provides their email, you might send "Checking our system..." while your workflow queries a database, then send the results without waiting for another response.

Send and Wait for Response

This operation pauses workflow execution until the user submits a response. The workflow remains in a waiting state, and execution only continues after the user interacts. Use this when you need:

• User confirmation before taking action
• Additional information to proceed
• Manual approval for sensitive operations
• Multi-step form collection

Example: Your AI agent analyzes a support ticket and proposes sending a $50 refund. Before executing, a Chat Node asks: "Should I process this $50 refund?" with Approval buttons. The workflow waits for the human operator to click Approve or Disapprove.

The key difference: "Send Message" is for one-way communication, while "Send and Wait for Response" creates true back-and-forth interaction.

Building Human-in-the-Loop Approval Workflows

One of the most valuable use cases for the Chat Node is building human-in-the-loop workflows that require manual approval before executing critical actions. This pattern is essential when working with AI agents, financial operations, or any automation that affects customers directly.

The Approval Response Type

When you set the Response Type to "Approval," n8n displays inline Approve/Disapprove buttons in the chat interface instead of a text input box. This creates a clear decision point and ensures users can't accidentally type invalid responses.

How to build an approval gate:

1. Add a Chat Node after the action you want to gate
2. Set operation to "Send and Wait for Response"
3. Choose "Approval" as the Response Type
4. Write your approval message with context: "AI suggests refunding $50 to customer@example.com for order #12345. Approve?"
5. Add an IF node after the Chat Node to check {{ $json.approval }}
6. Route approved actions to your execution nodes (Stripe refund, database update, etc.)
7. Route disapproved actions to a different path or end the workflow

This pattern ensures that AI-powered workflows remain under human control while still automating the research, analysis, and preparation steps.

5 Practical Chat Node Workflows You Can Build Today

1. AI Customer Support with Human Escalation

Build a support chatbot that handles common questions with AI but seamlessly hands off to humans when needed:

• Chat Trigger: User opens chat and asks a question
• OpenAI Node: AI analyzes the question and generates a response
• Chat Node (Send Message): Delivers the AI's answer
• Chat Node (Send and Wait - Approval): "Did this answer your question?"
• IF Node: If user clicks Disapprove, route to Slack notification to alert support team
• Chat Node (Send Message): "A human agent will help you shortly"

This workflow lets AI handle 80% of requests while ensuring frustrated customers get human attention.

2. Lead Qualification Bot

Qualify leads conversationally before adding them to your CRM with complete data:

• Chat Trigger: Visitor starts chat from your pricing page
• Chat Node (Send and Wait - Free Text): "What's your company name?"
• Chat Node (Send and Wait - Free Text): "How many team members?"
• Chat Node (Send and Wait - Free Text): "What's your biggest automation challenge?"
• HubSpot Node: Create contact with all collected data
• Chat Node (Send Message): "Thanks! Our sales team will reach out within 24 hours."

Each response is automatically saved and passed to your CRM, eliminating manual data entry.

3. Internal Refund Approval System

Create a chat interface for your support team to request refunds with automatic approval routing:

• Chat Trigger: Support agent opens internal refund tool
• Chat Node (Send and Wait): "Enter order ID"
• Shopify/Stripe Node: Look up order details
• Chat Node (Send Message): Display order summary
• Chat Node (Send and Wait - Approval): "Refund $127.50 to customer@example.com?"
• IF Node: Check approval status
• Stripe Node: Process refund if approved
• Google Sheets Node: Log refund in finance spreadsheet
• Chat Node (Send Message): "Refund processed. Confirmation: ref_abc123"

This replaces manual refund request emails with an instant, trackable chat interface.

4. AI Content Writer with Editor Approval

Generate blog post drafts with AI, then get human approval before publishing:

• Chat Trigger: Marketing manager enters blog topic
• OpenAI Node: Generate outline and draft
• Chat Node (Send Message): Display generated outline
• Chat Node (Send and Wait - Approval): "Proceed with full draft?"
• OpenAI Node: Generate complete article (only if approved)
• Chat Node (Send Message): Show full draft
• Chat Node (Send and Wait - Approval): "Publish to WordPress?"
• WordPress Node: Create draft post (only if approved)
• Chat Node (Send Message): "Draft created: [link]"

This workflow keeps humans in control while automating the time-consuming writing process.

5. Database Cleanup with Safety Gate

Build an admin tool that requires confirmation before deleting customer data:

• Chat Trigger: Admin opens cleanup tool
• Chat Node (Send and Wait): "Enter customer email to delete"
• Postgres Node: Query all associated records (orders, subscriptions, support tickets)
• Chat Node (Send Message): Display summary: "Found 3 orders, 1 active subscription, 12 support tickets"
• Chat Node (Send and Wait - Approval): "This will permanently delete all data for user@example.com. Are you sure?"
• IF Node: Only proceed if approved
• Multiple Postgres Nodes: Delete from all tables
• Slack Node: Log deletion to audit channel
• Chat Node (Send Message): "Deletion complete. Audit log: #12345"

This prevents accidental data deletion while still providing a fast interface for GDPR/compliance requests.

Chat Node Best Practices and Tips

Always provide context in approval messages: Instead of "Approve this action?" write "Refund $50 to jane@example.com for order #12345?" Reviewers shouldn't need to check other systems to make decisions.

Use "Send Message" for status updates: When your workflow is processing data or calling APIs, send status messages ("Analyzing your request...") so users know the workflow is working and haven't been disconnected.

Validate user input immediately: After a "Send and Wait for Response" that expects an email address or order ID, validate the format in the next node. If invalid, use another Chat Node to ask again rather than letting the workflow error.

Set reasonable timeouts: Chat workflows that wait for responses should have timeout handling. Use the Wait node with a branch that sends a "Are you still there?" message if no response arrives within your timeout period.

Consider embedding chat in your existing tools: While hosted chat is great for testing, production workflows often benefit from embedding the chat widget directly in your internal admin panels or customer-facing dashboards using the embedded chat option.

Log all approval decisions: For compliance and auditing, always log who approved or disapproved actions, when, and what context they saw. Use a Google Sheets or database node after each approval gate to maintain an audit trail.

Test your chat flows thoroughly: Unlike traditional workflows that run the same way every time, chat workflows have multiple possible paths based on user input. Test both the happy path and edge cases like users typing unexpected responses or clicking Disapprove when you expect Approve.

Use the Approval type for yes/no decisions: Free text responses work for collecting information, but for simple approve/disapprove decisions, always use the Approval response type. It's faster for users, prevents typos ("yes" vs "Yes" vs "y"), and provides cleaner data for your IF nodes.

The n8n Chat Node transforms automation from a set-it-and-forget-it tool into an interactive assistant that collaborates with humans. Whether you're building AI agents that need oversight, internal tools that require approval gates, or customer-facing chatbots that escalate to humans, the Chat Node gives you the flexibility to pause workflows exactly where human judgment adds the most value.

Teams running chat workflows on n8nautomation.cloud get reliable 24/7 uptime for their chat interfaces, automatic backups of conversation logs, and dedicated instances that handle multiple concurrent chat sessions without slowdowns—all starting at $15/month with no server management required.