1. Scope the pilot
Choose one class or intervention cohort with a ceiling of 30 learner seats. Capture year level, buyer contact, privacy owner, approved SchoolAdmin and intended start date.
School Evidence Pilot · Phase C
Run a school-ready maths intervention pilot.
Math.Foundation gives one Australian school a bounded path from diagnostic baseline to classroom intervention evidence and a documented rollout decision. The default evidence window is six instructional weeks for up to 30 learners.
2. Clear the trust gates
Complete buyer procurement/privacy review, prove the isolated persistence gate, confirm school-scoped consent, and provision the exact SchoolAdmin and teachers before learner activation.
3. Import learners and run diagnostic week
Import learners only after the hard gates are green. Students complete placement and live lesson probes while teachers review misconception flags and the first class heatmap and the baseline evidence packet is frozen.
4. Produce rollout evidence
After the six-week intervention window, close with observed mastery/gap evidence, support and incident evidence, and a renew, expand, revise, or stop decision.
Full offering readiness
Code and buyer documents are not the final proof. Live learner activation remains on HOLD until the hard institutional gates are satisfied, including R33-25.
Pilot classroomOne school, one cohort, maximum 30 learner seats
AuthorityExact product-owned SchoolAdmin, teacher, enrolment and school-consent boundaries
PersistenceR33-25 isolated migration, concurrency, restore and lifecycle proof required before live learners
ProcurementAustralian buyer-usable onboarding/purchase evidence is the active pilot gate
InternationalTracked separately as future expansion evidence; not an Australian pilot blocker
EvidenceBaseline, class gap evidence, close-out report and explicit rollout decision
Evidence contract
Safety thresholds are hard: exact school authority and consent, zero confirmed cross-tenant authority events, and zero unresolved critical privacy/security incidents at close-out. Learning improvement is measured against the target agreed before launch; Math.Foundation does not invent a universal efficacy percentage before real School #1 evidence exists.