Product brief

The primary user is a campaign reviewer deciding whether a specific creative version is ready for a simulated approver. Losing the connection between feedback, review and replacement assets creates an avoidable ambiguity: which version did the decision cover?

Decision: every comment, requested change, handoff and approval remains attached to its original asset/version. Every introduced replacement requires new review and a new exact-version approval. This conservative rule makes scope explainable in a small sample; classifying cosmetic changes as exempt was rejected because this prototype has no reliable content-difference policy.

Smallest observable outcome: a reviewer annotates the poster v1, requests a change, introduces v2, reviews it and hands it off; an approver confirms only v2. A prior v1 approval is visible as historical and never authorizes v2.

This is a visibly provisional product recommendation. Mo owns product/program direction, with AI assisting implementation and verification. Personal understanding and stakeholder acceptance require observation; software evidence does not establish either.

Non-goals: production DAM, login/access enforcement, real approval authority, invitations, messages, external services/Box, employer or customer media, live AI, payments and analytics. All Northstar people, creative and review decisions are original fictional samples.

Principal alternative: keep one approval on the asset and carry it forward. That reduces clicks but hides the identity of the reviewed media. Exact-version review costs an additional handoff but prevents implied replacement approval.

Proposed measures are correct version identification, requested-change clarity, approval-scope comprehension and recovery. The proposed human evaluation and release gates are in Validation.