# AWR-007-LAB1 — learner workspace

Build `Multi-region/Ad Automation Risk Decision Matrix v1` from visible synthetic evidence. The starter has no prefilled final decisions. Keep the Multi-region and Ad Automation axes independent, preserve the exact shipped fixture evidence, diagnose the four mutations, inspect `fixture://AWR007-MISSING-EVIDENCE-01` for missing decisive evidence, and recompute the transfer fixture.

Run: `python run_checks.py evidence/your_matrix.json`

`CONSISTENT` means only that the submitted matrix exactly binds to the declared shipped fixture and satisfies the published `CHECK_RULES.json`. It is not production infrastructure approval, permission to spend advertising money, or a hidden answer key.


Consistency note: each positive axis records an explicit operational/ownership-cost class. Ad Automation `USE` has a stricter bar than `PILOT` and requires shipped-fixture evidence that the simulation/recommendation progression stage was completed. The checker never invents this evidence.


The missing-evidence case intentionally omits `failover_owner`. With exact fixture binding preserved, a positive Multi-region state must be rejected; the checker does not choose an expected answer, so you must record an allowed non-success state that fits the evidence.