That check is arranged with the actual inserts, not just photographs, and I've kept the current description as appears reversible until tooling has tried the intended and unintended combinations under its inspection procedure.

Does the recipe currently identify an insert revision, or only the housing family?

Only the housing family, and tooling has now confirmed that both inserts mount and the intended one also mounts reversed; the controls and fixture owners are comparing identification and orientation checks against those actual combinations.

Make the current and proposed combinations explicit in that comparison. An ID check could recognise the correct insert while missing its wrong orientation. And a distinctive locating feature might distinguish orientation without establishing which family has been fitted.

Who owns the response when the observed insert or orientation disagrees with the selected recipe, including after an interrupted changeover?

Controls owns that response with the application assessor, and the draft leaves the job unaccepted after a mismatch or an incomplete changeover check; the fixture proposal adds orientation-specific location, but its physical combinations still need verification.

Who updates stores if that fixture proposal changes the insert references?

Chloe's stores question belongs with the release work. The operator needs the same identities on the insert, its storage location and the recipe record; three separately sensible naming systems could leave this just as confusing.

Our tooling coordinator owns those linked references, including stores; the physical review now confirms the proposed mounting rejects reversed seating, while the wrong-family insert still needs its separate identity check, so we haven't called the whole changeover validated.

Has a covering operator walked the incomplete-check and wrong-family cases without the author explaining the messages, before the combined design proceeds to application verification?