Louis Philippe · Van Heusen · Peter England · Allen Solly

Headswap at Production Scale for India's Biggest Menswear Brands

Headswap Pipeline · 6–8K Images/Month · Live on 4 Brands

6–8K

Images / Month

Brands Live

<90s

Per Image

AI Characters

Overview

Four national menswear brands of Aditya Birla Fashion and Retail (ABFRL) — Louis Philippe, Van Heusen, Peter England and Allen Solly — needed consistent model imagery across thousands of garments shot every month. 8M Studio's headswap pipeline, deployed inside production partner Studio 11 Productions in Bangalore, puts brand-owned AI characters on every garment shot.

The Challenge

Premium catalog shoots meant flying models in from abroad — high cost, a thin supply of the right faces, constant scheduling pain, and model usage rights that expire. The studio shoots ~3,500 garments a month; at that volume, those costs and dependencies compound fast.

Our Solution

8M Studio trained 16 brand-approved AI characters and built a ComfyUI pipeline that swaps them onto every garment shot — the clothing stays 100% real, only the model is AI. Version 1 ran fully on cloud GPUs: great quality, but slow and expensive per image. A custom caching method cut inference time in half with no quality drop, letting the pipeline move fully on-prem to the studio's own GPUs. The studio's own 'AI operator' runs it day to day; 8M Studio supports, trains and upgrades.

Results

Live on all four brands' product pages since June 2026, passing enterprise brand QC — skin texture, eye direction, hair consistency, grain matching. Current throughput is 6,000–8,000 images per month and growing, at under 90 seconds per image. In the brands' own words: the AI shots are looking better than the human-model ones.