AI Tutor day boundary, required-unavailable fallback, and submit-timestamp completion policy

DEC-0031 - AI Tutor day boundary, required-unavailable fallback, and submit-timestamp completion policy

Context

After locking reminder delivery behavior, the product still needs deterministic day-state semantics and dead-end protection when required content is unavailable.

Decision

Day boundary:

• AI Tutor daily required state uses user local timezone with day boundary at 00:00.

Required unavailable fallback:

• If today’s required item becomes unavailable, system auto-replaces it using existing A -> B -> C fallback ladder.
• System shows a short informational notice when replacement occurs.

Completion-day attribution:

• Daily required completion and day attribution use submit timestamp.

Decision Value

• Local 00:00 day boundary solves ambiguity on “today” state and keeps reminder/day-state consistent.
• Auto-replacement on unavailable required solves dead-end risk and protects completion continuity.
• Submit-timestamp attribution solves reporting/streak disputes near day boundary.

Rationale

These rules keep flow predictable for learners and implementation-light for operations. They reuse existing fallback logic, avoid adding new branches, and maintain one clear source for daily completion truth.