Automation has always been the secret weapon of productive teams, but in 2026, combining workflow automation platforms with large language models has unlocked an entirely new level of efficiency. If you have ever wished that your apps could talk to each other intelligently — reading emails, summarizing documents, generating responses, and updating your CRM without a single manual click — this tutorial is for you.
Why n8n + OpenAI Is the Automation Combo to Learn in 2026
n8n is an open-source, node-based workflow automation tool that lets you connect hundreds of services together without writing complex integration code. Unlike competitors that lock you into monthly subscriptions, n8n can be self-hosted for free, giving you full control over your data and workflows.
When you pair n8n with the OpenAI API, you gain access to GPT-4o and newer models that can understand context, generate natural language, classify information, and extract structured data from unstructured text. This combination turns simple “if-this-then-that” rules into genuinely intelligent processes that adapt and reason.
What You Will Build: An AI Email Triage and Response System
In this tutorial, you will build a workflow that automatically:
- Monitors your inbox for new emails
- Classifies each email by intent (support request, sales inquiry, spam, or general)
- Generates a draft response tailored to the email content
- Labels the email in Gmail and logs it to a Google Sheet
This is a real-world use case that freelancers, small business owners, and solo entrepreneurs use daily to save hours of manual email management.
Step 1: Set Up n8n
The fastest way to get started is using Docker. Run the following command to spin up n8n locally:
docker run -it --rm --name n8n -p 5678:5678 -v n8n_data:/home/node/.n8n n8nio/n8nOnce the container is running, open your browser and navigate to http://localhost:5678. You will see the n8n editor canvas where you can start building workflows visually.
If you prefer cloud hosting, n8n also offers a managed service at n8n.cloud with a generous free tier that handles up to 2,500 executions per month.
Step 2: Configure Your OpenAI Credentials
In the n8n editor, go to Settings → Credentials → Add Credential and search for “OpenAI.” You will need an API key from the OpenAI platform. Head to https://platform.openai.com/api-keys, generate a new key, and paste it into the credential configuration.
Make sure your OpenAI account has billing enabled and a small credit balance. The GPT-4o model costs approximately $2.50 per million input tokens and $10 per million output tokens as of mid-2026, which means classifying and drafting responses for hundreds of emails will cost only pennies per day.
Step 3: Build the Email Trigger Node
Click the + button on the canvas and add a Gmail Trigger node. Authenticate with your Google account and configure it to watch for new messages in your inbox. Set the polling interval to 2 minutes for near-real-time processing.
The trigger node will output the email subject, body, sender, and other metadata for each new message it detects.
Step 4: Add the AI Classification Node
Add an OpenAI node after the trigger. Configure it with the following settings:
- Model: gpt-4o
- Operation: Chat Completion
- System Message: “You are an email classification assistant. Classify the following email into exactly one category: SUPPORT, SALES, SPAM, or GENERAL. Respond with only the category name.”
- User Message: Reference the email subject and body from the previous node using n8n’s expression syntax:
{{$json.subject}} - {{$json.text}}
This node sends the email content to GPT-4o and receives a single-word classification back. It is fast, accurate, and costs less than $0.001 per classification.
Step 5: Generate Draft Responses Conditionally
Add an If node to branch the workflow based on the classification result. For SUPPORT and SALES emails, you want to generate a draft response. For SPAM, you can skip it. For GENERAL, you may want a brief acknowledgment.
For the SUPPORT and SALES branches, add another OpenAI node configured to generate a polite, helpful response draft:
System: You are a professional customer support agent. Write a concise, friendly reply to the following email. Keep it under 150 words. Do not make commitments about pricing or timelines unless the email provides that information.
User: {{$json.subject}} - {{$json.text}}This prompt ensures the AI generates appropriate drafts without overpromising. You can always review and tweak the drafts before sending, which brings us to the next step.
Step 6: Label in Gmail and Log to Google Sheets
Add a Gmail node (not the trigger — the action node) to apply a label based on the classification. Create labels like “AI-Support,” “AI-Sales,” “AI-Spam,” and “AI-General” in Gmail beforehand, then use an expression to select the correct label dynamically:
{{$json.category === 'SUPPORT' ? 'AI-Support' : $json.category === 'SALES' ? 'AI-Sales' : $json.category === 'SPAM' ? 'AI-Spam' : 'AI-General'}}Next, add a Google Sheets node to append a row to a tracking spreadsheet. Include columns for timestamp, sender, subject, classification, and the AI-generated draft. This log gives you a searchable record of every email processed by the system.
Step 7: Test and Activate
Before activating the workflow, click the Test button to run it with a sample email. Verify that the classification is correct and the draft response makes sense. Adjust prompts or branching logic as needed.
Once satisfied, click Activate in the top-right corner. Your workflow will now run automatically whenever a new email arrives.
Advanced Tips to Level Up
- Add sentiment analysis: Ask the AI to rate email sentiment (positive, neutral, negative) alongside classification, then escalate negative-sentiment support emails to a human agent via Slack notification.
- Auto-reply for simple cases: For SPAM and GENERAL classifications, configure n8n to send an auto-reply directly instead of just creating a draft.
- Multi-language support: Modify the system prompt to detect the email language and respond in the same language automatically.
- Cost monitoring: Use n8n’s execution logging to track OpenAI token usage per workflow run. Set up a monthly budget alert using a simple Google Sheets formula that sums token costs.
- Self-host for privacy: If you handle sensitive business emails, run n8n on a private VPS and consider using a self-hosted LLM like Ollama with Llama 3 for classification instead of the OpenAI API.
Conclusion
Building AI-powered automation workflows with n8n and OpenAI is no longer a futuristic experiment — it is a practical skill that freelancers and businesses are using right now to save time, reduce costs, and scale their operations without hiring additional staff. The email triage system we built in this tutorial processes hundreds of messages for pennies, responds faster than any human could, and keeps a perfect audit trail in your spreadsheet.
The beauty of this approach is its flexibility. Once you understand the pattern — trigger, classify with AI, take action — you can apply it to customer support tickets, social media mentions, form submissions, Slack messages, or any other data stream that flows through your business. Start small, iterate fast, and let the AI handle the repetitive work so you can focus on what actually matters.