Generative AI Integration for a Leading Clothing Retailer

Objective:

One of the UK’s largest online clothing retailers was looking at ways to reduce the need for workers to complete data input for product descriptions, which is often seen as a repetitive and time consuming task. We were chosen for our leading state-of-the-art technology that allows full product descriptions to be generated from scratch.

Challenge:

There is a thinking that generative AI could make several industries and jobs redundant. However, the intention of this service is not to eliminate the need for humans but to give them more time for more complex and skilled objectives, like writing more creatively about key niche products.

Solution:

The suppliers of the product provided a unique number and image. Our image detection model tagged and calculated all the attributes of the product, including its size, fit, material and care. It then passed this to a large language model which has been trained by experts so that it matches the clothing brand’s output structures and guidelines. This was pushed through a system called Fluid 4, where it was then uploaded onto the website. In most cases, there is no need for human involvement.
In this process, we also provided an AI confidence score. Every generation gets a score measuring the accuracy of the description. If it was below a certain value, around 90%, it will be reviewed by a human.
Following our successes, all product descriptions will be generated by Aria from November 2024.

“Generative AI has the ability to work alongside employees to complement their workload and remove the need for repetitive tasks. This not only saves time and increases profits, but it helps to empower employees who can instead prioritise higher-level duties. We’re delighted with these results and we can’t wait to help others across the retail sector in the coming years!”

Debonil Chowdhury, CEO and Co-founder of Aria