A shared checklist for product, engineering, compliance, and operations to decide what is validated, what is still risky, and what work remains before any real member interacts with the recommendation and enrollment experience.
Click any item's status chip to cycle Not started → Partial → Validated → Blocked. State is local to this demo session.
What we must know before building
Align on the member problem, the current journey, and the definition of ready before writing a line of production code.
Why: Avoid building a faster path to the wrong product.
Evidence: User research synthesis; personas (e.g., Margaret, the caregiver proxy); problem statement canvas.
Risk if skipped: Scope creep; features that impress but don't move enrollment quality.
Why: Baseline to measure improvement and identify handoff points.
Evidence: Journey map + baseline metrics (NPS, call handle time, conversion rate, error rate).
Risk if skipped: No credible ROI; cannot detect regression.
Why: Determines non-negotiable compliance gates.
Evidence: Regulatory matrix; TPMO, SOA, marketing, privacy, licensing requirements.
Risk if skipped: Fines, carrier contract termination, marketing suspension.
Why: Prevents duplicate infrastructure and reveals integration constraints.
Evidence: Systems inventory; data dictionaries; API contracts; licensing agreements.
Risk if skipped: Underestimated integration cost; data rights issues.
Why: Aligns the VP and team on the decision gate.
Evidence: Signed readiness criteria and exit thresholds (accuracy ≥ 95%, completion ≥ 60%, hallucination < 1%).
Risk if skipped: Subjective launch decisions; hidden disagreements.
Why: Shapes liability, disclaimers, and escalation paths.
Evidence: RACI matrix; legal review of liability language.
Risk if skipped: Ambiguous accountability when a recommendation is wrong.
documents/plans/production-readiness-rubric.md