How to Build AI Workflow Automations with Make.com: Step-by-Step 2026 Tutorial

AI workflow automation is revolutionizing how teams streamline operations, from intelligent document processing to automated decision-making. Make.com (formerly Integromat) remains a top choice for building robust, low-code AI-driven workflows in 2026, thanks to its deep integrations, intuitive visual builder, and flexible automation capabilities.

As we covered in our complete guide to low-code and no-code AI workflow automation, this area deserves a deeper look. In this tutorial, you'll learn—step by step—how to design, build, and deploy a practical AI workflow automation using Make.com, with hands-on examples and troubleshooting tips.

Prerequisites

• Make.com Account (Free or Pro, 2026 version)
• OpenAI API Key (or other AI provider; we'll use OpenAI GPT-4o in this tutorial)
• Google Workspace Account (for Gmail and Google Sheets integration)
• Basic familiarity with automation concepts (no coding required, but helpful)
• Optional: Node.js v20+ and curl (for API testing)

This tutorial assumes you want to automate the following scenario: When a new email arrives in Gmail, analyze its sentiment with OpenAI, and log the result to a Google Sheet.

Step 1: Create a New Scenario in Make.com

1. Log in to Make.com.
   Navigate to make.com and sign in.
2. Click Create a new scenario.
   Screenshot description: The 'Create a new scenario' button is located on the dashboard, top right corner.
3. Add the Gmail module.
   In the scenario canvas, click the + icon, search for "Gmail," and select Watch emails as the trigger.
4. Connect your Gmail account.
   Authorize Make.com to access your Gmail. Select the label or inbox you want to monitor.

Step 2: Parse Email Content

5. Add a Text module (optional, but recommended).
   This helps extract or clean the email body. Add the Text → Extract plain text module and map the Body field from the Gmail trigger.
   Screenshot description: The scenario canvas now has two modules: Gmail (Watch emails) and Text (Extract plain text), connected by an arrow.

Step 3: Integrate OpenAI for Sentiment Analysis

6. Add the OpenAI module.
   Click the + icon, search for "OpenAI," and select Create a completion (for GPT-4o or GPT-4 Turbo).
7. Configure the OpenAI module:
   • Connect your OpenAI account using your API key.
   • Set Model to gpt-4o or gpt-4-turbo.
   • In the Prompt field, enter:
     Analyze the sentiment of the following email. Reply only with 'Positive', 'Negative', or 'Neutral'.
Email: {{2.text}}
   • Map the output of the previous Text module to the Email variable.
   Screenshot description: The OpenAI module is configured with the correct model, prompt, and mapped input from the Text module.

Step 4: Log Results to Google Sheets

8. Add the Google Sheets module.
   Click +, search for "Google Sheets," and select Add a row.
9. Connect your Google Sheets account.
   Authorize Make.com to access your Google account.
10. Configure the module:
    • Select your spreadsheet and worksheet.
    • Map the following fields:
      • Date/Time: {{1.date}} (from Gmail)
      • From: {{1.from}}
      • Subject: {{1.subject}}
      • Sentiment: {{3.choices[0].text}} (from OpenAI)
    Screenshot description: Google Sheets module field mapping panel, with dynamic fields from Gmail and OpenAI mapped to columns.

Step 5: Test and Activate Your Scenario

11. Click Run once to test.
    Send a test email to your Gmail. Watch as the modules execute in sequence. Check for green checkmarks on each module.
12. Review the output in Google Sheets.
    Confirm that a new row appears with the correct details and sentiment.
13. Click Activate to enable your automation.
    Your workflow is now live!

Step 6: (Optional) Add Error Handling and Notifications

14. Add a Router module for branching.
    Use this to handle errors (e.g., if OpenAI returns an error or ambiguous result).
15. Add a Telegram or Email module.
    Configure it to send alerts if sentiment cannot be determined.

For more advanced workflow branching and security best practices, see our secure procurement AI workflow tutorial and this guide on shadow IT risks.

Sample OpenAI API Request (for Testing)

If you want to test the OpenAI integration outside of Make.com, use curl:

curl https://api.openai.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_OPENAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "messages": [
      {"role": "system", "content": "You are a sentiment analysis assistant."},
      {"role": "user", "content": "Analyze the sentiment of this email: Thank you for your help, I appreciate it."}
    ]
  }'
  

Common Issues & Troubleshooting

• Gmail module not triggering?
  Ensure your Gmail connection is active and that the label/inbox is correctly selected. Double-check account permissions.
• OpenAI module returns errors?
  Verify your API key is valid and has sufficient quota. Check that the prompt format matches the API's expectations.
• Google Sheets row not added?
  Make sure the spreadsheet and worksheet exist, and your account has write access. Confirm that all required fields are mapped.
• Rate limits or timeouts?
  Both Gmail and OpenAI have rate limits. Space out your tests and consider batching if needed.
• Data mapping issues?
  Use Make.com's Output panel to inspect data at each step. Remap fields if necessary.

Next Steps

• Expand your workflow: Add more AI steps (e.g., summarization, classification), integrate with Slack, Teams, or CRM systems.
• Explore advanced automations: Use routers, iterators, and error handlers for complex logic.
• Learn about platform selection and risk management: See our comparison of no-code vs. low-code AI platforms and shadow IT risks in no-code AI workflow environments.
• Discover more tools: For team-specific recommendations, see the best no-code AI workflow tools for small teams and our finance-focused roundup.
• Compliance & deployment: For regulated environments, see our AI workflow automation checklist for finance.
• Prompt engineering: Learn how business analysts can design custom AI workflows in our no-code prompt engineering tutorial.

For a broader strategic view, revisit The 2026 Guide to Low-Code and No-Code AI Workflow Automation—Platforms, Risks, and Roadmaps.