AI entitlement visibility and credit fairness in result policy

DEC-0058 - AI entitlement visibility and credit fairness in result policy

Context

Tier boundaries existed, but execution-level clarity was still needed for free/pro visibility depth, credit re-consumption on reopen, refund behavior on AI system failures, and mid-attempt entitlement drops.

Decision

Tier visibility:

• Free Writing/Speaking shows basic total score only.
• Pro remains basic-scoring only for Writing/Speaking.
• Detailed AI analysis remains locked unless entitlement supports advanced AI.

Credit fairness:

• Reopening an already-scored result does not consume additional AI credit.
• AI credit is charged once per scoring job/result.
• If AI scoring fails due to system error, consumed AI credit is auto-refunded.

Mid-attempt safety:

• If entitlement/credit becomes unavailable during attempt, submission is still allowed.
• At result, unavailable AI sections are shown as locked with upgrade/top-up entry.

Decision Value

• Prevents perceived double-charging and improves trust in AI credit model.
• Keeps learning flow uninterrupted even when entitlement changes mid-attempt.
• Makes free/pro boundaries explicit and consistent for users and support teams.

Rationale

These rules reduce support disputes, keep paywall behavior predictable, and preserve low-friction learning continuity.