Agent roles
Agent roles define what each component may do and, importantly, what it may not.
AI for Kids · Class 10 · Module 7 of 9
Automation with Human Control
Delegate the work. Never delegate the responsibility.
8 sessionsSessions 49–56
4 weeksTypical pace
AppliedLevel
Class 10Artificial Intelligence
The big idea
Students design automation that is powerful and controllable — clear agent roles, approval and audit trails, and failure-safe behaviour that degrades gracefully instead of collapsing.
Why it matters
This is the professional standard now emerging around agentic systems, and understanding it early is a genuine advantage.
Unpacked
What you will actually learn
Three core ideas, each taught with worked examples and then practised until it feels obvious.
Approval and audit
Approval and audit mean a record of what was done, by which component, and who authorised it.
Failure-safe design
Failure-safe design ensures that when something breaks the system stops safely rather than continuing incorrectly.
Did you know?
Safety-critical engineering assumes failure will happen and designs for it. Software that assumes success is the outlier, not the norm.
Common mix-up
“Logs are only for debugging.” Logs are the accountability record. Without them, nobody can establish what actually happened.
Session by session
Your 8-session journey
Sessions 49–56 of the 72-session year, at two one-hour sessions per week.
- 1
Warm-up and big picture
Where this module fits, what you will build, and a hands-on starter that gets everyone curious about agent roles.
- 2
Agent roles
Guided teaching on agent roles, worked through together with the teacher.
- 3
Agent roles — practice lab
Independent practice, small challenges and one deliberate mistake to diagnose.
- 4
Approval and audit
Guided teaching on approval and audit, building directly on the previous two sessions.
- 5
Approval and audit — practice lab
Applied tasks that combine agent roles and approval and audit in one piece of work.
- 6
Failure-safe design
Failure-safe design introduced and practised, completing the toolkit needed for the project.
- 7
Project build
Guided build session for the module project: Responsible automation blueprint.
- 8
Test, present and reflect
Finish, test against the checklist, present the work and explain the decisions behind it.
Project lab
Responsible automation blueprint
Every module ends with something the student built themselves and can demonstrate. This is the piece that goes into their portfolio and gets explained out loud at the end of session 56.
Try this at home
Take your automation and ask: if this ran wrong for a week unnoticed, what would the damage be? Design the alarm for that.
It is finished when
- It works from start to finish without breaking
- You can explain every part of it in your own words
- You tested it and improved at least one thing afterwards
- Someone else used or understood it without your help
By the end of this module
Students finishing Module 7 can:
- Explain and use agent roles without prompting
- Explain and use approval and audit without prompting
- Explain and use failure-safe design without prompting
- Build and finish responsible automation blueprint
- Test your own work and correct what you find
- Talk an adult through what you made and why
Word bank
The vocabulary introduced here, in plain language:
- audit trail
- A record of actions taken by a system.
- fail-safe
- Defaulting to a safe state when something fails.
- authorisation
- Permission required before an action proceeds.
Module challenge
Think you have got this?
0 XPLevel 1 · Curious Beginner
6 quick questions drawn from this module — vocabulary, the project you build, and a myth-or-fact round. Every wrong answer explains itself, so a mistake still teaches you something.
Your child's free trial is one message away.
Tell us your child’s class and what they enjoy. We will suggest the closest program fit—no pressure and no upfront payment.