Do-it-yourself game plans

Stage 1: How do you write a useful learner brief?

Describe observable starting skills and one achievable outcome—not an ambition to “master AI.” Complete the brief with the child before suppliers propose a programme.

Explore the comparison in this briefing

Age band

01 / 4 options

8–11

Suggested starting work
Sorting activities, pattern examples, adult-guided AI literacy
Evidence to look for
Can describe examples, mistakes and why a computer might get something wrong
Initial boundary
Prefer activities without open-ended personal chatbot use

Compare each option against the same criteria. The complete table and qualifications remain in the briefing.

Copy-and-use learner brief

Field What to write
Adult decision-maker Parent or guardian; coordinator’s authority
Learner Age band initially; exact age provided securely where eligibility requires it
Interests Two or three genuine interests, preferably in the learner’s own words
Starting evidence Tasks completed independently; tasks completed with help
Desired capability “Explain and modify a small classifier,” not “become an AI expert”
Proposed project One sentence, with a deliberately small first version
Available time Session availability and realistic between-session time
Teaching preferences Pace, communication style, accessibility needs and frustration triggers
Tool constraints Approved accounts; prohibited uploads; no public deployment initially
Household rules Parent-visible communication, recording preference and escalation contact
Budget Trial ceiling, block ceiling and approval rule for extras
Success evidence Explanation, modification, debugging and later recall
Stop conditions Unsafe conduct, ineligible tools, unauthorised charges or learner unwillingness

Share relevant learning accommodations with permission and on a need-to-know basis. The first enquiry does not need a medical history or an unredacted school assessment.

How do you check readiness without examining the child?

Try a friendly, approximately 20-minute baseline:

  1. Ask the learner to show something they previously made.
  2. Ask what they did, what someone else supplied and what AI generated.
  3. Invite one small change to the work.
  4. Ask how they would check whether the result is correct.
  5. Record what support was needed, without giving a grade.

If the child cannot yet make a small change, that is a scoping finding—not a failure. Begin with a simpler task or coding foundations.

What age-banded starting points should you consider?

The following are suggested project choices, not developmental standards. Demonstrated ability and account eligibility matter more than the band alone.

Age band Suggested starting work Evidence to look for Initial boundary
8–11 Sorting activities, pattern examples, adult-guided AI literacy Can describe examples, mistakes and why a computer might get something wrong Prefer activities without open-ended personal chatbot use
12–14 Small coding tasks, labelled examples and simple model evaluation Can change an input, predict a result and describe a test Treat a 12-year-old’s tool eligibility separately from a 13-year-old’s
15–17 Bounded model or AI application project Can justify design choices, inspect errors and document assistance No public launch by default; check every service’s minor-user rules
18 More independent project management if ready Can reproduce the workflow and take responsibility for operation Revisit consent, account ownership and parent access rather than assuming the minor arrangements continue

Sources and context

This page sets out the desk’s purchasing framework. Sources →