Evidence

What changed?

Descriptive comparisons with the sample and limitations visible.

Baseline · n=8Generic support

→

Adapted · n=9Human Intelligence applied

Message clarity+0.9 score3.2 → 4.1 · favorable

Clarification requests-0.9 count2.1 → 1.2 · favorable

Reported usefulness+18 percent58 → 76 · favorable

DESCRIPTIVE · NOT CAUSAL. The samples are descriptive and may differ in task mix, context, and exposure.