The Z-Image Enhancer options appear to have little influence on the result.

[sylvanmotenai-dot]

With the standard prompt the results obtained are very similar (regardless of whether the Enhancer options are enabled or not). Therefore, I tested including the SeedVarianceEnhancer node and many consistent variations. I'm just a newbie and I'm only starting out with ComfyUI and Z-Image, and I don't know if this is the ideal point to include this node, but since it worked, I'm reporting it here.

Insertion point of the Seed Variance Enhancer node (the rest is unchanged):

[martin-rizzo]

Happy New Year!
Thank you for sharing the idea of including that node! I must admit I'm not familiar with it, but from your screenshot it seems to be placed correctly within the workflow. The variations are noticeable, and I love how she always looks badass in all of them!
It's worth noting that the enhancer included in the default workflow doesn't modify the prompt itself or change the composition, but rather enhances the final image as a post-processing step. It adds more detail when the z-image generation comes out too flat or lacks sufficient texture.