Create an original, production-quality image-generation prompt for this brief: create a controlled set of visual ad variations where only one creative variable changes at a time. Inputs: {{product}}, {{base_scene}}, {{variables_to_test}}, {{brand_rules}}, {{aspect_ratio}}. Define the focal subject, environment, composition, camera or illustration perspective, lighting, material behavior, depth, color mood, and negative space. Preserve exact product, architecture, logo, packaging, UI, or person-reference details when supplied; do not redesign them unless asked. Produce six labeled variants with consistent product, framing, and typography space while changing only the specified hook element, setting, lighting, or audience cue. The final visual should look intentional at advertising or editorial quality, not like generic AI art. Avoid distorted anatomy, duplicate objects, fake text, clutter, plastic skin, inconsistent reflections, impossible structure, and excessive effects. If copy will be added later, keep a clean text-safe zone and do not render small body copy into the image.