Make.com Customer Support Automation: Handle Tickets Automatically

Support Automation Architecture

A complete support automation in Make.com:

Inbound channels: Email, live chat (Intercom), WhatsApp (Twilio), web form (webhook)

Processing: Claude AI classifies, analyses, and drafts responses

Output channels: Ticketing (Zendesk/Freshdesk), CRM (HubSpot), team notifications (Slack), direct response

Knowledge base: Airtable or Notion → Claude uses this to ground responses

Email Support Automation

Scenario: Gmail → AI classify → route

1. Gmail trigger: New email to support@
2. OpenAI or Claude HTTP: Classify

1. Parse JSON
2. Router: Branch by can_auto_resolve

- True → AI response path - False → Human routing path

AI response path:

1. Airtable: Search knowledge base for relevant articles
2. Claude HTTP: Draft specific response using knowledge base
3. Gmail: Create draft (human reviews) OR Send directly for clearly answerable questions
4. Zendesk: Log ticket as resolved

Human routing path:

1. Router (second): Branch by urgency + category

- critical → Slack #urgent-support + Zendesk P1 ticket - billing + frustrated → Billing team email + P2 ticket - technical → Engineering Slack + Zendesk technical queue

1. Zendesk: Create ticket with AI summary + category + urgency
2. Slack: Notify team

Live Chat Integration

Intercom + Make.com + Claude:

1. Intercom trigger (webhook): New conversation started
2. Claude HTTP: Generate first response

- System: "You are a friendly support agent. Keep responses under 150 words (chat context). Use knowledge base: {{knowledge_base}}" - User: "{{customer_message}}"

1. Intercom: Post bot reply to conversation
2. Track: If customer replies again → fire another Make.com webhook → Continue conversation

Handoff to human agent: In Intercom, configure bot to trigger "handoff" when:

• Customer asks for a human
• Third round of conversation (Claude can't resolve)
• Customer sentiment = frustrated

Handoff webhook → Make.com → Slack: "Customer needs human support: {{conversation link}}"

Multi-Channel Support

WhatsApp support (Twilio + Make.com):

1. Twilio webhook: inbound WhatsApp message
2. Make.com Data Store: get conversation history for this phone number
3. Claude HTTP: context-aware response with history
4. Twilio: send WhatsApp reply
5. Data Store: update conversation history

Unify across channels in Zendesk: All channels (email, chat, WhatsApp) create Zendesk tickets. The customer has one support history regardless of channel.

AI Knowledge Base Integration

Claude's responses are only as good as the context you give it. Ground responses in your documentation:

Airtable knowledge base setup: Create an Airtable base with:

• Article name
• Category (billing, technical, account, etc.)
• Content (full FAQ or doc text)

In your Make.com scenario:

1. Before Claude call: Airtable → Search Records (category matches ticket category)
2. Text Aggregator: combine matching records into one text block
3. Include in Claude system prompt: Knowledge base: {{aggregated_text}}

Fallback instruction in Claude prompt:

Support Analytics

Turn support data into product insights:

Weekly support digest:

1. Schedule: Monday morning
2. Zendesk: get all tickets from last week
3. Text Aggregator: compile ticket subjects and descriptions
4. Claude: "What are the top 5 recurring issues? What product changes would reduce ticket volume? What documentation gaps exist?"
5. Notion: create weekly support insight report
6. Slack: post to #product channel

CSAT tracking:

1. Zendesk webhook: ticket closed
2. Email: send CSAT survey (1-5 stars)
3. Survey response → Make.com → Airtable: log score + ticket + category
4. Monthly: Claude analyses trends → email report

Recommended Tools

• Make.com — Support automation backbone
• Claude API — Ticket classification and response drafting
• Zendesk — Ticketing platform with native Make integration
• Intercom — Live chat with webhook support
• Twilio — WhatsApp and SMS channels
• Airtable — Knowledge base storage