Video summary

Build & Sell n8n AI Agents & Automation (Full Detailed Course )

Main summary

Key takeaways

Technology

Technological Concepts & Course Focus (AI Agents + Automation with n8n)

  • The video is presented as a full, detailed course designed to take viewers from beginner to advanced in building and selling AI agents and automation workflows.
  • It emphasizes that you can build automation/AI agents without strong technical knowledge, gradually introducing custom code only where necessary.
  • The main platform used throughout is n8n (shown as “N10/n8n” in subtitles), alongside external AI/tools such as:
    • OpenAI / ChatGPT models
    • Google services (e.g., Gmail, Google Search API, Google Calendar)
    • Gmail
    • WhatsApp / Telegram
    • 11Labs
    • Additional APIs (via integrations)

Core Product Features / Workflow Building in n8n

1) “Agents” vs “Automation” (Key Conceptual Distinction)

  • Automation is framed as:
    • Trigger → Action → Result
    • The Action is mostly pre-defined and doesn’t reason dynamically.
  • AI Agents are framed as:
    • Trigger still starts the process, but the “Action” becomes an AI reasoning step
    • The agent can analyze inputs at runtime and choose different paths, such as routing an order to the correct department based on its contents.

2) Five Node Types Described in n8n

The course explains workflows conceptually through five node types:

  1. Trigger node
    • Starts the automation/agent (e.g., click, form submit, email/order received).
  2. Action node
    • Performs tasks (e.g., send email via Gmail, integrate WhatsApp/Telegram, operate Google Calendar).
  3. Utility node
    • Filters/transforms data (e.g., conditional routing using thresholds like order value).
  4. Code node
    • Custom JavaScript for parsing/formatting (e.g., split an address into city/postcode).
  5. Agent node
    • The “brain” using LLM logic (e.g., Gmail email triage; WhatsApp alerts for “urgent”/important emails).

Example AI Agents Built in the Course

Agent 1: “Smart Email Responder”

  • Goal: Monitor Gmail, classify incoming emails, and generate personalized replies.
  • Process shown:
    • Trigger: Gmail “new email” polling (e.g., every minute)
    • Text classifier: Categorize emails into groups like Orders vs Inquiries
    • LLM step (OpenAI chat model): Generate a draft reply using the classified content
    • Output handling: Save drafts to Gmail Drafts (not send directly) to reduce errors
  • Prompting approach:
    • A structured prompt requesting:
      • Subject line
      • Email body
  • Demonstrated results:
    • Drafts created for order confirmation and inquiry responses

Agent 2: “Mobile Accessories Inventory System”

  • Goal: Automate inventory management for mobile shops.
  • Data source: Airtable (also compared to Google Sheets)
  • Agent behavior:
    • Answers chat questions like: “What is the stock quantity of X?”
    • Searches the inventory table and responds using current quantities
    • Updates inventory via commands like:
      • “Update stock of X to 100”
      • voice/chat updates
  • n8n setup steps described:
    • Create an Airtable base (products with fields like price, quantity, etc.)
    • Create API tokens/credentials
    • In n8n, create tool/actions:
      • Search
      • Update
      • Delete
    • Use expressions (and optionally record IDs) to update the correct Airtable row
  • Demonstrated result:
    • Stock values updated and verified in Airtable

Agent 3: “Personal AI Assistant” (Voice Assistant / Jarvis-like)

  • Uses 11Labs for voice output.
  • Uses:
    • OpenAI (or similar) for chat reasoning
    • Google Search API for real-time internet information
    • V0 (v0.dev) to generate the app scaffold
  • Features described:
    • Wake-word style commands (e.g., “Jarvis turns on/off lights”)
    • Intelligent Q&A
    • Voice responses
    • Internet search for up-to-date facts via the Search API
  • Course emphasis:
    • Connect multiple APIs
    • Set environment variables

Tutorial Mechanics for n8n: Building Workflows Step-by-Step

Workflow Structure and Testing

The course repeatedly demonstrates:

  • Creating a project
  • Building a workflow
  • Adding nodes (trigger/action/utility/agent/code)
  • Using test/execution to validate results
  • Using test drafts (for email) and verification steps (e.g., checking Airtable after updates)

Webhooks + “Jarvis calling sub-agents” Architecture

For the Jarvis web/app experience, the course outlines an architecture:

  • The Jarvis interface sends requests via a webhook to n8n
  • n8n runs an LLM-based router/agent (Jarvis main agent)
  • It then calls specialized sub-tools/sub-agents, such as:
    • Calendar booking agent
    • Email agent
    • Content creator agent
    • (and more expandable agents)
  • It also explains:
    • Using APIs like Google Calendar for create/update/delete operations
    • Passing request data as JSON fields (e.g., transcript/user command text)
    • Returning a response back through the webhook so Jarvis can speak the result

Selling / Market Framing & Course Business Layer

  • The video frames AI agents as a business opportunity:
    • Claims very few people can provide agent/automation services
    • Mentions that people are making money by building and selling systems
  • Mentions a curriculum progression:
    • Build basic agents
    • Build lead generation / outreach automation
    • Build more advanced business workflows
    • Sell services and use intermediate/advanced sales modules

Pricing, Platform Cost, and Hosting Guidance

n8n Cloud Plan & Self-Hosting

  • Mentions n8n plans:
    • Example base plan around $20
    • Pro around $50
    • Notes 14-day trial behavior: agent access disappears after the trial unless subscribed
  • To avoid access loss, the course recommends self-hosting:
    • Uses Hostinger VPS
    • Suggests a VPS size (including RAM concerns like 4GB vs 8GB for “big agents”)
    • Chooses a location (e.g., Malaysia)
    • Explains that hosting means running automation 24 hours, like a dedicated server
  • Mentions coupon/discount guidance (including a coupon code shown in the flow)

Main Speakers / Sources Mentioned

  • Main speaker: Course instructor/host (unnamed in subtitles; uses “I” and demonstrates screens)
  • Primary tool/source references:
    • n8n (spelled as “N10/N8N” in subtitles)
    • OpenAI / ChatGPT models
    • Gmail, Google Calendar, Google Search API
    • Airtable (and a comparison to Google Sheets)
    • 11Labs (voice generation)
    • V0 (v0.dev) (app code generator)
    • Hosting provider: Hostinger

Original video