Skip to Content
View all events

n8n AI Bootcamp: Building AI Workflows and Agents From Scratch

Published by O'Reilly Media, Inc.

Intermediate content levelIntermediate

Automate Real-World Tasks with AI Agents, LLMs, and No-Code Workflows

Course Outcomes:

  • Understand the fundamentals of n8n and how to create powerful no-code/low-code automations.
  • Integrate state-of-the-art LLMs such as Gemini or OpenAI’s APIs.
  • Switch between different AI providers with minimal configuration effort.
  • Construct agentic workflows using memory, custom tools, and Model Context Protocol (MCP).
  • Address key considerations for security, monitoring, and scalable deployment.
  • Explore front-end customization and on-premise hosting options.

Course Description

n8n is one of the most powerful workflow automation tools, enabling both developers and non-developers to build scalable automations with minimal code. In this hands-on bootcamp, you’ll learn how to create AI-powered workflows using LLMs like Gemini or OpenAI’s APIs – and take your automation skills to the next level.

On the first day, you'll dive into the fundamentals of n8n and build a complete AI-powered workflow from scratch based on a real-world use case. You’ll learn how to go from building a simple prototype all the way to production. On Day 2, you’ll explore more advanced concepts and evolve your project into a fully agentic chatbot – adding autonomous decision-making and custom tool use – before enhancing it with Retrieval-Augmented Generation (RAG) for dynamic knowledge retrieval. You'll also explore further optimization strategies, monitoring options, and deployment considerations for scaling your agent into production environments.

You’ll walk away with practical skills to integrate LLMs, design robust AI automation logic, and deploy intelligent agents in real-world scenarios. Whether you're streamlining internal operations or building AI-driven products, this course equips you to start fast—and scale confidently.

What you’ll learn and how you can apply it

  • Gain hands-on experience building AI-powered workflows in n8n.
  • Integrate LLMs like Gemini or OpenAI via API to create intelligent automations.
  • Apply prompt design and data handling techniques within n8n flows.
  • Build agentic workflows with memory, conditional logic, and tool use.
  • Leverage custom tools and Model Context Protocol (MCP) to orchestrate more complex agent behavior.
  • Understand deployment strategies, monitoring options, and on-prem hosting.
  • Apply best practices for security, reliability, and maintainability in AI workflows.

This live event is for you because...

  • You are a developer or automation engineer who wants to build LLM-powered workflows.
  • You are a data and operations professional seeking to automate repetitive AI tasks.
  • You are a product manager or technical lead exploring internal AI agents and tools.
  • You are an AI enthusiast looking to combine LLMs with low-code tools like n8n.
  • You have a basic understanding of APIs and want to build real-world AI agents.

Prerequisites

  • Basic familiarity with APIs and JSON data formats.
  • Some experience with workflow or automation tools (e.g., Zapier, Make, n8n) is helpful but not required.
  • n8n Cloud (starter plan, or free trial) or self-hosted n8n and access to Gemini, Claude, OpenAI, DeepSeek, Mistral, or Ollama API key.
  • A modern web browser and stable internet connection.

Course Setup:

Recommended Preparation

Recommended Follow-Up

Schedule

The time frames are only estimates and may vary according to how the class is progressing.

### Day 1: Your First AI Workflows in n8n (4h)

  • Welcome and Course Introduction (15 minutes)
  • Presentation: Overview of the course, schedule, and outcomes
  • Demo: What you’ll build

n8n Fundamentals (45 min)

  • Presentation: How does n8n work? Introduction to nodes, flows, triggers, and data handling
  • Exercise: Getting started with n8n and building your first basic automation
  • Q&A

Integrating LLMs into n8n Workflows (45 min)

  • Overview: OpenAI node vs LLM Chain node vs HTTP Request node
  • API credentials setup
  • Demo: Connecting to Gemini or OpenAI APIs
  • Prompt structure and context handling
  • Monitoring token usage
  • Chaining LLM outputs across nodes
  • Exercise: Build a simple AI-powered text classifier with an LLM chain
  • Q&A
  • BREAK

Evaluations in n8n (30 min)

  • Presentations: What are evals and why are they important
  • Demo: Overview n8n’s built-in Evaluation module
  • Exercise: Create a light evaluation set for your text classifier

Build Phase 1: Q&A Support Chatbot (60 min)

  • Real-world use case: Internal IT Helpdesk chatbot
  • Demo: Workflow logic, error handling, and chaining nodes
  • Exercise: Launch your first custom AI chatbot

Deployment (30 min)

  • Deployment strategies (cloud/self-hosted)
  • Considerations before you go live:
  • Security & Privacy considerations
  • Logging + error alerts
  • Versioning strategies
  • Demo: Customizing your chatbot in n8n
  • Exercise: Deploy your customized chatbot

Wrap-Up Day 1 (15 min)

  • Recap what we’ve covered
  • Outlook Day 2
  • Closing Q&A

### Day 2: Advanced n8n: AI Agents, RAG, and MCP (4h)

Welcome to Day 2 (15 min)

  • Recap learnings from Day 1
  • What to expect today
  • Q&A

Agentic Workflows and Design Principles (45 min)

  • Presentation: What makes a workflow agentic?
  • Memory, autonomy, and decision-making
  • Tool-use inside n8n for AI Agents
  • Demo: Building a simple AI agent in n8n

Build Phase 2: Agentic Support Chatbot (45 min)

  • Use Case Introduction: IT Helpdesk Agent that can answer questions, resolve issues and create/update tickets (tbc)
  • Demo: Refactor IT Helpdesk Chatbot to be agentic
  • Add tool use: Log conversation using a ticket ID (e.g., Github)
  • Update and track ticket status ("open", "closed", etc.)
  • Add memory, retry logic, and fallbacks
  • Handle API calls, logs, and context
  • BREAK

Introduction to Retrieval-Augmented Generation (45 min)

  • Quick intro to RAG: What it is and when to use it
  • Concepts: Vector embeddings, similarity search, and knowledge retrieval
  • Exercise: Build the Embedding Flow
  • Setting up your vector database (Pinecone)
  • Creating vector embeddings from internal documents in n8n
  • Uploading the embeddings to Pinecone

Build Phase 3: Agentic Support Chatbot with Internal Knowledge Access (45 min)

  • Demo: Enhancing the Support Agent with RAG
  • Build the Retrieval Flow: Retrieving chunks and enriching metadata
  • Adding the Knowledge Tool: Have the support agent retrieve internal knowledge using RAG
  • Exercise: Try the agentic support chatbot
  • Compare different models and prompt
  • See where it breaks
  • Break

MCP and AI Agents in Production (30 min)

  • Adding more tools: How MCP fits in
  • Understanding Model Context Protocol (MCP)
  • Examples of MCP tools / servers
  • Demo: MCP in n8n
  • Presentation: Moving AI agents into production
  • Recap: What we built (and what’s still missing)
  • Optimization strategies for AI Agents (Tools, performance, UI)
  • Monitoring & evaluation strategies for AI agents
  • AI Agent Production Checklist

Wrap-Up (15 min)

  • Recap full transformation: Prompt-only bot → Agent → RAG Helpdesk Agent
  • What’s next: multi-agent systems
  • Share additional learning resources and templates
  • Closing Q&A

Your Instructor

  • Tobias Zwingmann

    Tobias Zwingmann is an experienced data scientist with a strong business background. He has more than 15 years of professional experience in a corporate setting, where he has been responsible for building out data science use cases and developing a company-wide data strategy. He is also a cofounder of the German AI startup RAPYD.AI and is on a mission to help companies adopt machine learning and artificial intelligence faster while achieving meaningful business impact.

    linkedinXsearch

Skill covered

Generative AI