A UK outlet fashion retailer whose stock changed faster than its team could write category copy or photograph products. In two weeks we built two AI workflows: one that generates content for hundreds of categories at once, and one that turns in-house photos into clean product shots.
This UK retailer sells outlet fashion, and outlet stock turns over fast. Products come and go, and the collections change with them. Every collection needs a category page with copy worth reading, and every product needs a description and photos good enough to sell it.
All of that was manual. Category content was written one page at a time. Product photos were taken in-house and uploaded one by one. The team couldn't keep pace with the stock.
We built two AI workflows in two weeks to take that work off their hands.
The first workflow generates category content in batches. The team can run it across 500 categories at once, and for each one it produces:
It can also be set up for different types of content, from general copy to FAQs and USPs, so each category gets the mix it needs.
Volume was the easy part. The hard part was making the content worth publishing: copy that reads properly, doesn't hedge, and helps someone decide what to buy. We grounded every output in the site's own knowledge base, its USPs and product details, so the copy says things that are true of this retailer rather than things you could say about any fashion site. Each batch then gets a humanising pass before it reaches review.
The second workflow handles images. The team uploads a photo of a garment taken in-house, and the AI recreates that exact garment as a clean product shot, no model, ready for the product page. The garment has to look identical to the real one. A customer who receives something different from the photo sends it back.
Running hundreds of images through an AI model gets expensive quickly, so we built in quality checks and batch processing to keep the cost per image down.
Nothing goes live without a person checking it. When a description or an image isn't right, the reviewer reruns only those selected rows or images, not the whole batch. Review stays quick, and fixing a bad output costs very little.
Category copy and product photos used to be produced one at a time. Now the team runs them in batches and spends its time reviewing instead of producing.
Both workflows came out of our AI workflows practice, specifically our workflow automation and integration work.
Every question we get asked on first calls. Answered in writing — decide before you book.
We take five new clients per quarter. Apply for a free audit and we'll tell you what we'd do first.