Premiata — AI lifestyle shoot
Generating lifestyle photos from a product packshot using Adobe Firefly and GPT Image 2 — a proof-of-concept for fashion e-commerce.
Context
A packshot as the starting point.
An exercise to see how much you can get from a packshot without a photo shoot. I took Premiata sneaker images — standard white-background studio shots — and ran a complete workflow from product reference to a full set of lifestyle photos.
Packshot sourced from Modivo.pl
Packshot sourced from Modivo.pl
Workflow
Prompt → reference → generate → retouch.
I started by writing prompts with Claude’s help. Each scene had its own brief: outfit style, location type, lighting character and mood. I wanted images that read like location photography in a Polish city — cool, natural light, concrete architecture.
Both packshots were attached as visual references in Adobe Firefly (GPT Image 2 model). The model accurately read the shoe’s shape and colour palette — the Premiata logo appears consistently across every generated shot.
The first results from the generator were slightly too warm and saturated for an urban street feel. A few minutes of colour and exposure correction was enough to make the images read more like outdoor photography than AI output. Below — a comparison of the first output and the corrected version:
Results
Five scenes, one product.
The generated scenes covered a range of lifestyle situations: full-body shot on a city street, dynamic movement crossing a pedestrian crossing, a couple on steps, a model leaning against a concrete wall, and a close-up of the shoe being tied.
Ad Creative
From photos to a finished banner.
Using the generated images I put together an example category banner — a collection tile with brand logo and CTA, ready for use in an online store or digital campaign.
Takeaways
AI speeds up the shoot, it doesn’t replace the process.
Precise prompting and strong reference images were the key to a working workflow. GPT Image 2 handles product detail consistency across generations very well — the logo, material texture and colour palette of the shoe remain coherent throughout the whole set. That makes the workflow scalable: the same packshot can be placed into dozens of scenes without a studio or a crew.
A complete set of lifestyle photos was ready in under an hour instead of a few days of production.