Make.com OpenAI Integration

Native OpenAI Modules in Make.com

Make.com includes these native OpenAI modules:

Create a Completion: Text generation using GPT-4o and other models. The core AI text processing module.

Create an Image: Generate images using DALL-E 3. Returns a URL to the generated image.

Create Transcription: Transcribe audio to text using Whisper.

Create Translation: Translate audio to English using Whisper.

Analyze Images: Pass an image URL and a question to GPT-4 Vision for image analysis.

Create Embeddings: Generate vector embeddings for semantic search applications.

All modules require only your OpenAI API key in Make.com's credential manager — no API request configuration needed.

Text Generation Workflows

Scenario: AI-powered support response drafting

1. Zendesk trigger: New support ticket created
2. HTTP Request: Fetch relevant documentation from knowledge base
3. OpenAI Create Completion:

- Model: gpt-4o - System: "You are a helpful support agent for [company]. Use the provided documentation to answer customer questions. Be concise, friendly, and accurate." - User: "Customer question: {{1.ticket_body}}\n\nRelevant documentation:\n{{2.docs_content}}"

1. Zendesk Add Internal Comment: Write the AI draft as an internal comment for the agent to review
2. Slack alert: Notify support agent — "AI draft ready for ticket #[number]"

The support agent opens the ticket, reviews the AI draft, personalises if needed, and sends. Average handle time drops significantly.

Audio Transcription Pipeline

Scenario: Voice note to CRM update

Sales reps often record quick voice notes after calls ("just talked to John, he's moving forward, wants a proposal by Friday"). Transcribe and process these automatically:

1. Trigger: New audio file in a specific Google Drive folder (reps share their voice notes there)
2. Google Drive Download: Download the audio file
3. OpenAI Create Transcription: Transcribe using Whisper
4. [Claude via HTTP]: "Extract from this sales call note: (1) contact name, (2) company, (3) key update, (4) next actions with dates. Return JSON."
5. HubSpot Search Contact: Find by name/company
6. HubSpot Create Note: Add the structured note to the contact record
7. HubSpot Create Tasks: For each next action identified

Voice notes become structured CRM data automatically.

Image Analysis Automation

Scenario: Social media content moderation

1. Webhook trigger: New user-submitted image (from your platform)
2. OpenAI Analyze Image: "Does this image contain any of the following: nudity, violence, hate symbols, spam content, or anything inappropriate for a professional platform? Return JSON: {'compliant': true/false, 'flags': [], 'confidence': 0-100}"
3. Router: Compliant vs. non-compliant

- Non-compliant → Hold image for review + notify moderation team - Compliant → Approve and publish

Scenario: Receipt and document data extraction

1. Gmail trigger: New email with attachment (from finance@ address)
2. Gmail Download Attachment
3. OpenAI Analyze Image: "Extract all text and data from this image. Then provide: vendor name, invoice number, date, total amount. Return JSON."
4. QuickBooks/Xero: Create bill with extracted data

Image Generation for Content

Scenario: Automated product image variants

1. Airtable trigger: New product record with status "Needs Marketing Image"
2. OpenAI Create Completion: "Write a detailed DALL-E image prompt for a product marketing photo for: {{1.productName}} - {{1.productDescription}}. The prompt should describe: a professional studio photo, clean white background, product centered, high contrast."
3. OpenAI Create Image: Use the generated prompt with DALL-E 3
4. Google Drive Upload: Save the generated image
5. Airtable Update: Link the image URL to the product record + set status to "Image Ready"

Combining OpenAI and Claude

Many sophisticated Make.com workflows use both OpenAI and Claude:

• OpenAI Whisper for transcription → Claude HTTP for structured extraction (Claude handles complex reasoning better)
• OpenAI GPT-4o for quick classifications (cheaper) → Claude for nuanced analysis that requires deeper reasoning
• OpenAI DALL-E for image generation → Claude HTTP for generating the image prompts and captions

Using both gives you the best capabilities from each model. OpenAI's native Make.com integration makes it fast to set up; Claude's HTTP integration requires more configuration but gives access to Claude's unique strengths.

Recommended Tools

• Make.com — Native OpenAI integration
• OpenAI API — GPT-4o, Whisper, DALL-E 3 access
• Claude API — Complex reasoning via HTTP module
• HubSpot — CRM destination for AI-processed data
• Google Drive — File storage for AI-generated assets
• Zendesk — Support platform for AI-drafted responses