Octimo Product blueprint
Proposed v0.1 Planning + prototype stage
Style guide Open prototype
Olympiad-first · responsive web

Build the learning loop
before the feature list.

Octimo turns exam preparation into purposeful play for learners aged roughly 8–14, while giving adults calm, auditable control. This blueprint connects the experience, product scope, system plan, and release gates needed to reach one reliable MVP.

Explore the delivery runway See prototype coverage
Audience
Product, design, engineering, content, and QA
Single job
Move from prototype to pilot without losing safety or learning intent
North star
Weekly Learning Achievers
01 · Experience

One brand, two working modes

Students need a clear next move in under ten seconds. Admins need to see consequence, ownership, and history before they act.

Student Portal Playful focus

“What should I learn next?”

One mission leads. Progress supports it. Rewards arrive after learning evidence, never before it.

  • Start, resume, and finish without distraction
  • Understand why an answer worked
  • See growth without public identity exposure
Enter Student Portal
Admin Portal Calm consequence

“What needs attention?”

Work is prioritised by impact. Risky changes show scope, validation, and an audit trail before they take effect.

  • Author, review, version, and publish safely
  • Trace scoring, access, and reward decisions
  • Control AI with default-deny capability gates
  • Keep basic learning independent of AI providers
Enter Admin Portal
Learning before engagement Child-safe by default Deterministic core AI never blocks basics Explainable outcomes Extensible curriculum
02 · MVP scope

Ship the dependable core; prepare the edges

P0 proves the complete learning loop. P1 deepens engagement and operations. P2 expands the business and delivery model without rewriting the foundation.

Capability MVP · P0 Next · P1 Later · P2
Learning Lessons, practice, quizzes, challenges, mocks Richer activities and sections Proctored workflows
Content Author, review, version, publish Advanced formats and imports External publishers
Motivation XP, levels, badges, streaks, safe rewards Missions, teams, seasonal leagues Curated reward marketplace
Insight Observed progress and operations Deterministic recommendations Validated adaptive models
AI + plans Kill switches and entitlement foundation Bounded pilot and adult plan control Plan-specific suite and billing
03 · Prototype coverage

What can be reviewed today

The current HTML makes fifteen high-value product surfaces reviewable. They remain illustrative concepts—not proof of API, permissions, persistence, or production accessibility.

STUDENT

Learning journey

  • HomeMission, goal, path previewOpen
  • LearnResponsive Octo-pathOpen
  • Practice setupPurpose and policy before startOpen
  • Practice playerAutosave and teaching feedbackOpen
  • Review and submitFinal status and lock confirmationOpen
  • ResultsEvidence, reward, next actionOpen
  • Answer reviewImmutable read-only releaseOpen
  • ProgressQualified evidence and historyOpen
  • Planned nextSign-in, lesson, compete, inbox, profile, and failure states
ADMIN

Operating surface

  • OverviewAttention, activity, healthOpen
  • PeopleScoped access issuesOpen
  • Question editorAuthoring and student previewOpen
  • Question reviewImmutable version evidenceOpen
  • Assessment builderPolicy, audience, and preflightOpen
  • Import resultsDry-run and row errorsOpen
  • AI controlsFail-closed access resolutionOpen
  • Design systemTokens, states, portal contrastOpen
  • Planned nextPrograms, content, moderation, reports, settings, and audit
04 · Delivery runway

Six gates from policy to expansion

This is an ordered sequence, not a dated progress report. Each phase ends in proof that the next investment is safe to make.

  1. 0

    Policy + foundation

    Decide

    Lock audience, consent, taxonomy, scoring, reward ledger, retention, audit, analytics, and the capability model.

    Exit gateApproved decisions and schema support one complete vertical slice.
  2. 1

    Core vertical slice

    Prove

    Build access, people, curriculum, authoring, Home, attempt autosave, deterministic scoring, results, and basic progress.

    Exit gateAn admin publishes one quiz; an eligible learner completes and reviews it end to end.
  3. 2

    MVP pilot

    Harden

    Add lessons, game systems, safe competition, moderation, reporting, imports, analytics, accessibility, and dormant AI controls.

    Exit gateAcceptance suite passes with reviewed pilot content and no release-blocking privacy or security issue.
  4. 3

    Engagement + operations

    Learn

    Add deterministic recommendations, missions, richer formats, guardian links, quiet hours, quality analytics, and seasons.

    Exit gateLearning and retention improve without adverse wellbeing or safety signals.
  5. 4

    Controlled AI pilot

    Optional

    Pilot bounded hints for an approved cohort with limits, moderation, reporting, cost controls, and a kill switch. Provider calls stay server-only and optional.

    Exit gateZero gate bypasses, approved safety and accuracy, controlled cost, evidence of learning benefit, and a complete core journey with AI off or unavailable.
  6. 5

    Plans + expansion

    Scale

    Add adult or institutional subscriptions, plan-specific capabilities, more programs, localisation, and later native apps.

    Exit gateEntitlements reconcile with account state and no purchase pressure reaches a child surface.
05 · Decision lock

Twelve choices block implementation lock

Safe defaults preserve momentum, but these choices change the operating model or architecture enough to require an owner.

12 open
01Audience + accounts 3 choices
  1. Confirm exact student age range, launch countries, and operating model.
  2. Choose admin-provisioned-only or verified guardian-created MVP accounts.
  3. Choose single-platform launch or immediate school tenancy.
02Curriculum + assessment 3 choices
  1. Approve Olympiad taxonomy, scoring formats, and negative marking.
  2. Confirm question types beyond single-select and true/false.
  3. Define author–reviewer separation, import scale, licensing, and duplicate rules.
03Safety + motivation 3 choices
  1. Choose dormant AI foundations or a bounded launch pilot.
  2. Define leaderboard opt-in/out and what counts as a safe cohort.
  3. Decide whether virtual currency is spendable in MVP and only on what.
04Data + expansion 3 choices
  1. Approve retention and erasure by data, audit, analytics, AI, and export class.
  2. Set guardian experience timing and future subscription ownership.
  3. Confirm launch languages, offline expectations, and validation scope.
06 · Release quality

The pilot opens only when trust is testable

Acceptance is organised around the harm or failure it prevents, not around team boundaries.

PROVISIONAL PILOT SCORECARD

Evidence that earns the next investment

Recalibrate after pilot
Activation≥ 70%First activity within 24 hours of first sign-in
Practice completion≥ 80%Started practice activities completed
Observed learning+10 ppAccuracy after three practices in one topic
Week-four retention≥ 30%Activated learners completing an activity in week four
Publishing speed< 10 minMedian quiz assembly after content is ready
Trust failures0Missing results/rewards, identity exposure, or unauthorised AI calls

Attempts + rewards

  • Reconnect restores every acknowledged answer.
  • Server time submits an expired attempt exactly once.
  • The same snapshot always produces the same result.
  • Retries grant no duplicate result, XP, badge, or notice.

Access + privacy

  • Students cannot reach admin or ineligible content.
  • Peer surfaces reveal approved aliases and avatars only.
  • Suspension revokes access and preserves authorised history.
  • Every failed AI gate denies before a provider call.
  • AI disablement or provider failure never blocks the basic student flow.

Content + versions

  • Invalid or inaccessible content cannot be published.
  • Published edits never alter active or completed attempts.
  • Answer keys stay hidden until the release rule allows them.
  • Reports identify the exact content version.

Access needs + audit

  • Core flows work with keyboard, zoom, and screen reader.
  • Motion and sound preferences preserve meaning.
  • Metrics reconcile to authoritative operational records.
  • High-impact changes record actor, target, time, and reason.
VALIDATION BEFORE BUILD LOCK

Test the risky moments, not just the happy screens

Learner navigation, interrupted quizzes, content publication, AI access resolution, leaderboard emotion, and a child best-interests review form the minimum research plan.

  1. 1 Learner comprehension by intended age band
  2. 2 Keyboard, touch, assistive tech, and reconnect
  3. 3 Draft → review → publish → correction
  4. 4 Default deny, expiry, kill switch, and quota
  5. 5 Competition comprehension and emotional response
  6. 6 Privacy and child best-interests impact review
07 · Source of truth

Read the implementation detail

This HTML is the review-friendly map. The Markdown specifications remain authoritative and carry requirement IDs, invariants, and references.

PRD Product requirementsScope, roles, P0–P2 requirements, metrics, acceptance, roadmap UX UI/UX specificationNavigation, screens, states, accessibility, safety, validation ARC Technical architectureBoundaries, runtime, access, events, AI gateway, operations DB Database designVersioned content, attempts, ledgers, capabilities, retention DBP Database structure planMigration slices, persistence gaps, read models, verification gates SCR Screen mockup planImportant-screen coverage, required states, contracts, delivery order