Make.com AI Automation

Make.com's AI Integration Options

Make.com connects to AI services in two ways:

Native integrations: Make.com has built-in modules for OpenAI (chat completions, image generation, Whisper transcription), Google AI (Gemini), and Cohere. These are pre-built with no API configuration required beyond your API key.

HTTP module for any AI API: The HTTP module sends any REST API request — use it for Claude, custom AI endpoints, Hugging Face models, and any AI service not in the native module library.

Both approaches work well. Use native modules when available for convenience; use HTTP for services without native modules (including Claude, which doesn't have a native Make.com module but is accessible via HTTP).

Connecting Claude to Make.com

The HTTP module configuration for Claude:

URL: https://api.anthropic.com/v1/messages Method: POST Headers:

• x-api-key: [your Anthropic API key]
• anthropic-version: 2023-06-01
• content-type: application/json

Body (JSON):

The response contains content[0].text — the AI's output. Map this to subsequent modules.

Lead Research and Enrichment

Scenario: Enrich new HubSpot contacts with AI research

1. HubSpot trigger: Watch for new contacts
2. HTTP Request: Fetch the company website (if domain is available)
3. Claude via HTTP: "Based on this company website content, write a 2-sentence description of what they do and who they sell to: [website content]"
4. Apollo.io HTTP: Fetch company data by domain (employee count, industry, funding)
5. Claude via HTTP: "Given this company profile, score this lead 0-100 for fit with our ICP [describe your ICP]. Return JSON: {'score': N, 'reason': '...'}"
6. HubSpot Update Contact: Write description, company data, ICP score back to the contact
7. Router: Route based on score — high priority creates a task; low priority moves to nurture

This scenario runs for every new HubSpot contact, so reps never see an unenriched lead.

Email Classification and Routing

Scenario: AI inbox triage for support email

1. Gmail/Outlook trigger: New email received in support inbox
2. Claude via HTTP: "Classify this email into one of: billing, technical, product-feedback, partnership, spam. Also estimate urgency: urgent, normal, low. Return JSON: {'category': '...', 'urgency': '...', 'summary': '...'}"
3. Router module: Branch by category
4. Each branch:

- Billing → Create Stripe/billing ticket - Technical → Create Zendesk ticket with high priority - Product feedback → Create Airtable record in feedback table - Partnership → Create HubSpot deal in "Inbound Partnerships" stage - Spam → Archive

Result: The support inbox self-triages. The right team sees the right emails. Nothing gets missed.

Content Generation Pipeline

Scenario: Blog post to social media repurposing

1. RSS trigger: New blog post published (watch your blog's RSS feed)
2. HTTP Request: Fetch the full article text
3. Claude via HTTP (parallel routes):

- LinkedIn post: "Write a professional LinkedIn post (150 words) promoting this article: [article text]" - Twitter thread: "Write a 5-tweet thread on the key points from this article: [article text]" - Email subject line: "Write 5 subject line options for an email promoting this article"

1. Buffer/Later module: Schedule the LinkedIn post for 10 AM Tuesday
2. Twitter module: Post the thread
3. Google Sheet: Log all generated content with article URL

Document Processing Workflows

Scenario: Invoice processing

1. Gmail trigger: New email with invoice attachment
2. Gmail Download Attachment
3. OpenAI module (Analyze Image or Vision): Extract structured data from the invoice — vendor, invoice number, date, line items, total
4. JSON Parser
5. QuickBooks/Xero module: Create a bill in accounting software
6. Google Sheets: Log to expense tracking spreadsheet
7. Slack notification: Alert finance team with extracted details

Recommended Tools

• Make.com — Scenario platform with HTTP module for AI APIs
• Claude API — Best for reasoning-heavy tasks (classification, analysis, writing)
• OpenAI API — Native Make.com integration; Whisper for transcription
• HubSpot — CRM trigger and action modules
• Apollo.io — Lead enrichment data API
• Airtable — Flexible data store for AI outputs

“Make.com's HTTP module makes any AI API a first-class automation participant. The platform's visual branching makes AI-conditional routing straightforward to build and easy to understand.”