Highlights
- Build three working workflows in one day
- Create a chat agent that answers questions from your own data, with memory and tools
- Clean messy data with exact rules, with no AI needed where rules do the job
- Add a guardrail that catches jailbreak attempts, and send serious issues to a person
- Prototype and demo one of your own tasks
- Learn the five building blocks that transfer to any automation tool
- Take home a delegate workbook
Course Details
Module 1: How AI agents work
- From using AI to building with it
- The five building blocks: trigger, data, steps, rules, guardrails
- Where AI judgement helps, and where exact rules are better
Module 2: Setting up your n8n workspace
- Importing data into data tables
- Creating tables and connecting your AI credentials
Module 3: Cleaning messy data with rules
- Pulling rows from a data table
- Removing duplicates and checking values with conditions
- Routing clean and rejected rows to separate tables
Module 4: Building a course advisor agent
- Setting up a chat trigger and an AI agent
- Writing a system message that keeps the agent on task
- Adding a chat model, memory and a data tool
- Testing the agent and seeing what breaks without memory
Module 5: Using AI safely
- Checking text for jailbreak attempts
- Extracting sentiment and follow-up flags from comments
- Escalating complaints and accessibility issues to a person
Module 6: Your own task: map, build, demo
- Mapping a real task across trigger, data, rules, AI step and guardrail
- Building a first working path with sample data
- A two-minute demo: the problem, one run, and what it would need for real use
Module 7: Next steps
- Rebuilding a lab on your own
- Checking data protection, naming an owner and testing before real use
Who should attend
- Operations, admin and process owners who want to automate repetitive work
- L&D, HR and service teams looking at practical AI use cases
- Analysts and team leads who want to prototype without waiting for IT
- Public sector and regulated teams who need AI with proper guardrails
- Developers who want a fast, visual way to understand agent design
Feedback
4.8 out of 5 average
"A lot in this course and suited to Python Data professionals - very fast paced but the instructor was careful to keep everyone engaged - well done Nick"
Hugh Last, Data Scientist (forensics) , May 2026