case study

FINE CHAOS: a full range of packshots, from photos the brand already had.

Copenhagen · Independent fashion brand · AW26 pre-release · fine-chaos.com

Packshot Studio produced the AW26 pre-release e-commerce packshots for FINE CHAOS, an independent Copenhagen fashion brand. 18 garments were delivered as clean on-white, ghost-mannequin hero images — generated by AI from the on-model photos the brand already had, and each one reviewed by a human against the real garment before delivery. As of July 2026, 19 product pages on fine-chaos.com lead with a packshot produced this way. No new shoot, no sample shipping.

What was delivered

18 garments delivered as on-white ghost-mannequin heroes
19 product pages on fine-chaos.com led by these packshots (July 2026)
3 angles — front, side, back — where the product page needed them
10/18 passed human review on the first generation; 8 needed revision passes
100% of delivered images individually reviewed against the real garment
0 new photo shoots, studio bookings, or samples shipped

Source material: the brand's existing on-model campaign and e-commerce photography — plus, for one garment below, phone photos taken in the store.

Before and after — three garments from the range

Left: the on-model photo the brand already had. Right: the delivered packshot that now leads the product page. Both are shown unedited at web resolution.

Heavy Jersey Skirt — the existing on-model photo used as source Before · existing on-model photo
Heavy Jersey Skirt — the delivered AI-generated ghost-mannequin packshot After · delivered packshot
Heavy Jersey Skirt. Layered draped jersey with leather eyelet straps — the strap-and-buckle construction had to survive the transformation intact. See it live on fine-chaos.com
Hooded Leather Jacket — the existing on-model photo used as source Before · existing on-model photo
Hooded Leather Jacket — the delivered AI-generated ghost-mannequin packshot After · delivered packshot
Hooded Leather Jacket. The review checked material read: the garment is smooth faux leather, and a version whose surface read as grained cowhide was caught and corrected before the current hero shipped. See it live on fine-chaos.com
Laced Hoodie — the existing on-model photo used as source Before · existing on-model photo
Laced Hoodie — the delivered AI-generated ghost-mannequin packshot After · delivered packshot
Laced Hoodie. Criss-cross leather lacing across the shoulders and sleeves, reproduced from the on-model source. See it live on fine-chaos.com

It also works from a phone photo

One pair of satin wide-leg trousers had no usable photography at all — only phone photos of the sample on a mannequin, taken in the store. From those, three ghost-mannequin angles were generated and reviewed. The delivered files are shown unedited.

Satin wide-leg trousers — the phone photo of the sample in the store, used as source
Source · phone photo, in the store
Satin wide-leg trousers — delivered AI-generated packshot, front
Delivered · front
Satin wide-leg trousers — delivered AI-generated packshot, side
Delivered · side
Satin wide-leg trousers — delivered AI-generated packshot, back
Delivered · back

How the run actually went

Each garment started from the source images the brand already had. The generation step produced a candidate packshot; a human reviewer then compared it against the real garment — silhouette, construction, colour, material, and any text or hardware — before anything was delivered or published.

Ten of the eighteen garments passed that review on the first generation. The other eight were rejected and re-run, some more than once. The rejections are the informative part: a coated denim that read as leather, a colour that drifted because a spec sheet disagreed with the photo, construction details the model invented that the real garment doesn't have. None of those reached the webshop — catching them is what the review step is for.

The first review wasn't the last word either. A second, independent review pass over the delivered set later caught further drift on several garments — a skirt whose layered drape had been flattened, a colourway that had shifted — and those heroes were re-run and replaced on the webshop as well. The packshots shown on this page are the versions that survived both passes.

Two patterns from the run, for anyone evaluating this kind of work: woven and embroidered garment text reproduced reliably from a good source crop, while large printed graphics were the hardest case. And detail that exists in the source — distressing, eyelets, hardware — carried through well; the failure mode to watch for is invention or omission, not reproduction. That is what the against-the-garment review checks.

Frequently asked questions

What did Packshot Studio deliver for FINE CHAOS?
Packshot Studio produced the AW26 pre-release e-commerce packshots for FINE CHAOS: 18 garments delivered as clean on-white, ghost-mannequin hero images, generated by AI from on-model photos the brand already had. As of July 2026, 19 product pages on fine-chaos.com lead with a packshot produced this way. No new photo shoot took place and no samples were shipped.
Is FINE CHAOS a real client?
FINE CHAOS is a real, independent Copenhagen fashion brand — and it is also our founder’s own label. Ludvig Isaksen founded both companies, and this range is where the Packshot Studio process was built and first used in production. We name it because we own it and can show the work with full permission; we state the relationship openly so you can weigh the case study accordingly.
Are the FINE CHAOS packshots AI-generated?
Yes. Every packshot in this case study was generated by AI from the brand’s existing on-model photography and then reviewed by a human against the real garment before delivery. The same disclosure appears on the FINE CHAOS work itself and across this site.
Did every image pass review on the first try?
No. Of the 18 garments in the run, 10 passed human review on the first generation and 8 needed one or more revision passes before delivery. A second, independent review pass later caught further drift on several delivered heroes, and those were re-run and replaced as well. Rejections were for things like colour drift, invented construction details, or material misreads — which is exactly what the per-image human review exists to catch. We publish these numbers because a review step that never rejects anything would not mean much.
Where can I see the results live?
On fine-chaos.com. Product pages such as the Heavy Jersey Skirt, the Hooded Leather Jacket, and the Laced Hoodie lead with the delivered packshots, with the brand’s original on-model photography alongside them.

See it on your own products, free.

Send a few images and we'll produce a reviewed sample set — from the images you already have, no shoot required.

Get a free sample

AI-generated product photography. Every image is created by AI from your existing images and reviewed before delivery. FINE CHAOS imagery shown with permission; the on-model source photographs are the brand's own photography and are not AI-generated.