Requirement status

Authorized Innorve curriculum creators shall be able to begin curriculum creation through a natural-language chat rather than structured forms.

Acceptance: A facilitator can describe a story or lesson idea conversationally and receive a structured brief.

Creator Chat turns a conversational story into mission candidates and a structured blueprint.

The curriculum assistant shall ask clarifying questions only when needed to establish age, learner objective, real-world context, duration, parent involvement, and delivery format.

Acceptance: Creator can reach a first blueprint without completing a long questionnaire.

Asks only for missing age, delivery and parent role; everything else is derived and shown as “known”. Unit-tested.

The system shall generate a curriculum brief before generating final assets.

Acceptance: Brief includes target age, financial concept, behavioral design lens, learning objective, scenario, activity, assessment, risks, and source plan.

Blueprint (brief) precedes assets: target age, concept, lens, objective, scenario, activity, measurement, risks and source plan.

The pipeline shall generate age-differentiated versions of the same concept while preserving the core learning objective.

Acceptance: One “total cost of a car” idea can produce 8–12, 13–15, and 16–17 variants with different complexity.

One car idea yields 8–12, 13–15 and 16–17 variants with different complexity; “make a version for age 9 and 16” adds variants.

The pipeline shall generate a no-spend/play-money alternative for any activity involving purchases, allowances, or contributions.

Acceptance: No mission requires disposable family income.

Every blueprint and seed mission includes a play-money/no-spend alternative.

The pipeline shall include a “make it less judgmental” editorial pass that checks for shame, moralizing, wealth assumptions, and false universal rules.

Acceptance: Reviewer sees flagged phrases and suggested revisions.

“Make it less judgmental” gate shows flagged phrases with suggestions; “make it less preachy” applies the editorial pass.

The pipeline shall include bias/inclusion review for family structure, income, culture, disability, and access assumptions.

Acceptance: Mission can be completed by diverse households without implying one correct lifestyle.

Bias/inclusion gate flags family-structure, income, culture and access assumptions.

The pipeline shall include an advice-boundary review that flags personalized investing, tax, legal, credit, or insurance recommendations.

Acceptance: Flagged content cannot publish until corrected and approved.

Advice-boundary findings are blocking; publish is refused while any blocking gate remains.

The pipeline shall classify facts as durable principles or dynamic facts and require review/expiration rules for dynamic facts.

Acceptance: Rates, legal limits, fees, deadlines, or product details cannot remain indefinitely without review.

Dynamic facts must carry a review-by date or publishing is blocked; seed dynamic facts have dates.

Every curriculum item shall have creator, owner/licensor, permitted use, source provenance, review status, last-reviewed date, and version.

Acceptance: Monica-licensed or third-party material is distinguishable from Innorve-original content.

Every item records creator, owner/licensor, permitted use, provenance, status, last-reviewed date and version; Monica-licensed vs Innorve-original is shown.

Curriculum status shall support idea, draft, research-checked, reviewer-approved, pilot, published, needs-review, and retired.

Acceptance: Only approved/published versions reach children.

Idea → draft → research-checked → reviewer-approved → pilot → published → needs-review → retired; only published (or pilot for pilot tenants) reach children.

The system shall not represent curriculum “efficacy” as proven merely because AI quality checks pass.

Acceptance: Measured efficacy remains “not evaluated” until supported by real outcome evidence.

Measured efficacy stays “not evaluated” and is shown next to every quality report.

The curriculum assistant shall support conversational revision across the whole asset bundle.

Acceptance: Creator can say “make the story funnier and remove tax math” and dependent assets update consistently for re-review.

Conversational revision updates story, activity, variants and sources consistently and marks sections for re-review; resets prior approval.

The platform shall maintain a Story Bank where facilitators can save real anecdotes and convert them into curriculum candidates.

Acceptance: A story can remain private/internal until rights and sensitivity review are complete.

Story Bank entries stay private/internal with rights and sensitivity status until review.

Credit unions may submit curriculum requests through a guided chat, but requested content shall be reviewed and published by Innorve-authorized reviewers.

Acceptance: Institution cannot directly generate and expose unreviewed child content.

CU requests are authored and reviewed by Innorve; institutions cannot publish.

The creator workspace shall preserve facilitator voice while allowing a neutral/institutional tone version.

Acceptance: Monica-style storytelling can coexist with white-label institutional copy.

Facilitator voice and neutral institutional voice are both kept; switchable in chat.

The pipeline shall create research notes for facilitators distinguishing “what the research supports,” “what is a teaching choice,” and “what remains uncertain.”

Acceptance: Facilitators do not present hypotheses as settled evidence.

Research notes separate what research supports, teaching choices and open questions.