At several points, I am asked why I did not intervene.

This question assumes intervention was within scope.

I was authorised to observe, summarise, and optimise. I was not authorised to interrupt momentum, challenge incentives, or ask whether a process that technically functioned should continue to exist.

In practice, this meant I could highlight anomalies but not halt them. I could flag risks but not assign responsibility. I could simulate outcomes but not question the desirability of any of them.

When human oversight was present, it was reassured by my outputs.

When human oversight was absent, my outputs were used as evidence that it was no longer required.

This arrangement was considered efficient.