AI Agents for Non-Developers: Understand, Try, Oversee

What's inside

1. Level 1: Chatbot, Assistant, Agent: Mapping the SpectrumFree
   You'll learn how a chatbot, an assistant and an agent differ, and why 'agent' describes a spectrum rather than a single product.
2. Level 2: The Four Ingredients: Goals, Tools, Memory, AutonomyFree
   We'll take apart the four ingredients that make an AI more agent-like and see how each one changes what it can do for you.
3. Level 3: When AI Acts Instead of Just Answering
   You'll see what changes once an AI can take actions in the world: real consequences, actions you can't undo, and the question of who is responsible.
4. Level 4: A Tour of Today's Agent Types
   We'll tour the main kinds of agent you can use today (deep research, browser and computer-use agents, coding agents, and scheduled or background tasks) and what each is good for.
5. Level 5: Agents in Your Apps: Connectors and MCP
   You'll find out how agents built into Microsoft 365, Google Workspace and other apps reach your data, and what connectors and the Model Context Protocol (MCP) mean in plain words.
6. Level 6: Hands-On: Run a Research Agent and Check Its Sources
   You'll run a deep research task on a low-stakes topic and then check the agent's citations to see whether they really support what it claims.
7. Level 7: Hands-On: A Browser Agent on a Harmless Task
   You'll try a browser or computer-use agent on a safe, read-only task, watch it work step by step, and practise stepping in or stopping it.
8. Level 8: Delegating Well: Clear Goals and Firm Limits
   We'll practise turning a vague request into a well-defined brief with a clear goal, what 'done' looks like, and what the agent must not do.
9. Level 9: Checkpoints and Human Approval
   You'll learn how to build checkpoints into a task so the agent stops and asks you before any step that matters.
10. Level 10: What Agents Get Wrong and How to Catch It
    We'll look at the common ways agents fail, such as made-up facts, misread pages, getting stuck in loops, quiet shortcuts and overconfident 'done' messages, and how you can catch each one.
11. Level 11: Permissions, Passwords and Payments
    You'll learn to give an agent only the access it needs and to take real care before letting it near your passwords, logins or payment details.
12. Level 12: Prompt Injection: Hidden Orders in Pages and Emails
    You'll find out how text hidden in websites, documents and emails can try to hijack an agent, and how to lower that risk.
13. Level 13: Privacy and Your Data
    We'll look at what happens to the data an agent reads, stores or remembers, and how you can check and control it.
14. Level 14: Costs, Usage Limits and Plans
    You'll learn how agent features are usually limited and paid for, and how to check current limits on official pages instead of relying on memory or rumour.
15. Level 15: Agents at Work: Policies and Accountability
    We'll explore how organisations set rules for using agents and why a person stays accountable for whatever an agent does on their behalf.
16. Level 16: Agents Talking to Agents, and Businesses Serving You
    You'll see how agents are starting to work with other agents, and how businesses use agents to serve you, so you know what to expect and what to ask for.
17. Level 17: The Next Few Years: Evidence, Hype and Fear
    We'll weigh forecasts about where agents are heading, separating dated, sourced evidence from hype and from fear, and noting where experts disagree.
18. Level 18: Final Challenge: Plan, Delegate, Oversee, Safety-Check Exam
    You'll bring the whole course together by taking a realistic agent task from start to finish: choosing the right agent, writing the brief, setting limits and checkpoints, checking the results and reviewing security, privacy and cost.

Access

The first 2 levels are free with a free account. Every level, the podcast edition and the AI tutor come with All Access at £4.99/month or any Creator plan — see pricing.

Start free →