AI PBL Bootcamp · Free · Three Mondays from 26 October
Launch

Onboard the external partner or client

An outside expert or client is invited without shared expectations about purpose, student role, deliverable, schedule, boundaries and feedback.

AI does

Preparation, gap detection and communication drafts.

You do

Vet the partner, protect students, avoid tokenism and maintain the relationship.

1

Research

Why this matters, and what good looks like

Partner onboarding means agreeing how an already confirmed partner and the school will work together. PBLWorks and HQPBL offer professional frameworks for authentic work and audiences. They do not verify a partner or demonstrate that this onboarding pack works. England's DfE guidance concerns local AI data protection, not a complete safeguarding protocol. Safeguarding means the school's procedures for protecting students. No source establishes AI benefit here. The sequence below is an authorial process choice; human agreements remain essential.

  • PBLWorks connects authenticity with real-world contexts, tools, standards or impact; clarify the actual project connection.
  • HQPBL describes sharing work with experts during development and at conclusion; agree the partner's actual feedback role.
  • DfE requires local planning and data-protection advice; keep personal data outside this workflow.

Where the evidence comes from

Strong for authentic audiences and community engagement; AI benefit unvalidated.

2

Workflow

Brief it, steer it, check it
  1. Establish the confirmed relationship

    Use one already confirmed partner and one bounded interaction. Bring your own product brief, operations plan and student pack, optionally from D03, D08 and L02. Equivalent teacher documents work. Keep vetting records and personal details offline.

    Prompt 1 · Relationship inventory
    CONTEXT
    Teacher-verified project and partner facts, source IDs/versions, confirmation record and school procedure summary without personal data: [materials].
    REQUEST
    Inventory purpose, mutual value, student role, deliverable, schedule, roles, interaction boundaries, feedback and safeguarding. Identify unresolved decisions for a teacher-partner conversation.
    QUALITY BAR
    Use only supplied facts. Never source a speculative partner or invent quotes, availability, commitments or endorsement. If partner confirmation or human safeguarding confirmation is missing, return HOLD with no onboarding plan. Missing individual facts remain HOLD. Keep names/contact details, private vetting records and student data outside AI. No grading or outreach.
    FORMAT
    If blocked: HOLD, missing evidence and human action only. Otherwise table: Field | Supplied fact | Teacher source ID/version | Status. Status: SUPPLIED or HOLD. End with Human questions, separating unresolved decisions from confirmed facts.
  2. Prepare the bounded discussion

    Verify the inventory yourself. Prepare a short discussion with the confirmed partner. Role options are proposals until both parties accept them. Include school safeguarding approval as a separate prerequisite; partner agreement cannot replace it.

    Prompt 2 · Discussion pack
    CONTEXT
    Exact step 1 output: [inventory].
    Teacher's explicit inventory approval/corrections, original fact record and approved school boundaries: [decisions].
    REQUEST
    Draft a mutual-value brief, short meeting agenda, student interaction protocol, feedback role and safeguarding checklist for discussion. Where choices remain, label them PROPOSAL and ask a question for mutual decision.
    QUALITY BAR
    No draft without teacher approval, confirmed partner and human-confirmed school safeguarding boundaries. Never invent availability, quotes, endorsements, obligations or policy. Every consequential fact needs teacher source ID/version. Do not describe proposals as agreements. Preserve student inquiry and ownership; no AI grading. No personal data or outreach.
    FORMAT
    If blocked: HOLD, missing evidence and human action only. Otherwise headings: Mutual-value brief; Meeting agenda; Student interaction protocol; Feedback role; Safeguarding checklist; Open decisions; Fact trace. Trace columns: Claim | Teacher source ID/version | SUPPLIED or PROPOSAL. Mark DISCUSSION DRAFT, NOT AGREED.
  3. Record mutual decisions with the partner

    Hold the short teacher-partner conversation and check roles and safeguards together. Record explicit decisions and school approval outside AI. Supply only a nonidentifying summary. If agreement needs more than 60 active minutes overall, stop and re-scope.

    Prompt 3 · Agreed revision
    CONTEXT
    Exact step 2 output selected by teacher: [draft].
    Original approved facts and school boundaries: [facts].
    Explicit teacher-partner decisions, exact edits, school safeguarding approval and source versions: [review].
    REQUEST
    Apply only supplied decisions and edits to all five pack components. Carry unresolved matters forward visibly; never complete them by assumption.
    QUALITY BAR
    Require explicit mutual approval of roles, commitments and interaction safeguards plus school approval. Missing or contradictory approval means HOLD with no agreed pack. Preserve source IDs/versions; never infer availability or endorsement. Student work remains student-owned. No personal data, grading, outreach or new commitments.
    FORMAT
    If blocked: HOLD, missing evidence and human action only. Otherwise headings: Mutual-value brief; Meeting agenda; Student interaction protocol; Feedback role; Safeguarding checklist; Decision trace. Trace columns: Decision | Teacher source ID/version | Applied or HOLD. Mark RECORDED HUMAN AGREEMENT, TEACHER CHECK REQUIRED.
  4. Compare the selected agreement with originals

    Select the exact revised pack and independently check it against the recorded agreements and school procedures. Audit discrepancies before storing it locally. AI cannot certify vetting or permission, and this method makes no contact or commitment.

    Prompt 4 · Agreement comparison
    CONTEXT
    Exact step 3 output selected by teacher: [selected].
    Approved original facts, mutual decisions and school approval record: [facts].
    Teacher's explicit selection and independent source checks: [selection].
    REQUEST
    Compare all five pack components with supplied records. Check mutual value, student ownership, roles, schedule, feedback, interaction limits and safeguarding. Identify unsupported commitments or missing approval.
    QUALITY BAR
    Do not rewrite or approve. Quote exact selected text for discrepancies. Unknown or stale facts remain HOLD; never invent partner quotes, availability or endorsement. Separate recorded human decisions from independently unverifiable real-world conditions. No personal data, grading or outreach.
    FORMAT
    Table: Check | Exact selected text | Teacher source ID/version | Finding | Human action. Findings: MATCH or HOLD. End with Operational status: HUMAN RELEASE REQUIRED; no outreach or commitment made. List untested checks explicitly.

Check before you use it

  • Confirm the real partner using your own records; never treat a draft as availability or endorsement.
  • Match each commitment and schedule detail to a current teacher source and version.
  • Obtain explicit teacher-partner agreement on roles, feedback and student interaction boundaries.
  • Check school safeguarding approval independently; mutual agreement alone is insufficient.
  • Keep student ownership visible and personal details, consent records and vetting evidence outside AI.
  • Stop for missing approvals; do not send outreach or make commitments from the draft.

Stop rule

Never invent a partner, quote, availability or endorsement; stop until commitments and safeguarding are confirmed by humans.

3

Reuse it

So the next one takes minutes
Suitable

Build a skill

Save this as a reusable instruction you can run on any project.

Stable drafting and comparison steps can be reused with fresh teacher materials and explicit human checks.

Not suitable

Build an agent

Chain the steps, with a checkpoint where you approve.

This bounded relationship task needs direct human decisions; recurring autonomous communication is outside scope.

Help us improve this method

Anonymous, and it takes under a minute.

Did you get something usable?
How useful was this method?
1 = wasted my time · 5 = I'll use it again
Where did it break down?

Please don't include names, school names or anything about individual students.

No account, no email, nothing stored about you.

Use cases are starting points with one perspective. Disagree with the process, that's expected. Make it yours.

Co-funded by the European Union under Erasmus+ KA210-SCH. No student data. No teacher material stored.