Brand consistency in fashion imagery depends on the same recognisable model recurring across the collection. Generic from-scratch AI fights you on this — the face drifts subtly between generations, and by SKU 47 you are showing a different person. Here is the practical workflow for locking a single AI fashion model identity across a 100-SKU drop.
Why identity drifts by default
Diffusion models are stochastic. Even with an identical prompt, identical seed, and identical settings, the output face shifts a few millimetres at every generation. The eye colour drifts. The hair gets a touch wavier. Scale this to 100 generations and the face on shot 1 is recognisably not the face on shot 100 to a careful viewer.
Brand teams often do not notice on a single shot. They notice when the catalog goes live and someone in marketing says “is that the same model on every product?”. Drift is invisible at the SKU level and obvious at the catalog level.
Approach 1: lock to an Apiway preset model
Apiway's White Studio ships ~50 female and ~10 male preset AI fashion models. Each preset is a stable identity that produces a consistent face across generations — the engineering work to lock the identity is done at the preset level rather than at the prompt level.
The recipe: pick one preset at the start of the catalog work and use it for every SKU. Same model, same aspect ratio, same framing baseline. Variation comes from the garment, not the model.
Approach 2: upload a real model reference
Stronger move when brand identity matters. Photograph a real model in your preferred light and pose. Upload that image to White Studio as the model reference. Apiway uses the upload as the identity anchor across every subsequent generation.
This is a Hollywood-anchor pattern at the catalog level — the face is real, the garment is the AI overlay. The 100-SKU catalog now has a face that is genuinely specific to the brand, not a stock AI model.
Approach 3: lock a creator photo set
For lifestyle and ad creative, pick one set from the marketplace and use it across the campaign. The same creator carries the brand recognition in every shot. Brands often pair this with the approach 2 above — one face for catalog, one creator for lifestyle, both consistent.
How to staff and manage the workflow
Document the model choice as a brand asset. Store the preset name or upload file alongside the brand colours and fonts in the same brand-system folder. Onboard producers explicitly — the model choice is a brand decision, not a per-SKU decision.
For 100 SKUs, the workflow runs in batches. Group SKUs by framing (mid-shot, close-up, full-body) and run each batch with the same model + framing combo. Roughly one operator-day total for 100 SKUs once the workflow is locked in.
Real cost and time at 100 SKUs
At one credit per shot and four shots per SKU on average, a 100-SKU catalog runs about 400 credits, or roughly $4. Operator time: one to two days for a clean first pass. (Pricing recap: one credit equals one cent.)
Lock identity before SKU 1
Pick the model approach now, before the catalog work starts. Switching identity mid-catalog produces visible drift and the team has to redo earlier shots. Open a free account and run two test SKUs against your chosen model preset before committing.