Posted on September 18, 2017 by staff AI 10,000 times faster than a human


Retail sales aggregator has launched a new machine learning platform to classify and categorise over one million products on its website, at a speed that could change the face of traditional merchandising.

For example, a task that would take a team of 10 four years to complete can be done using its machine learning in just under eight hours.

LovetheSales works with over 700 multi-national retailers, processing millions of products a day. The company brings together sales items from across ecommerce sites to allow consumers to shop every sale in one place.

With this comes the challenge of managing between 10 to 50 times the number of items even the largest of retailers have in their product catalogue. Enter Minerva, the AI system developed by LovetheSales.

Minerva’s role is to understand what each product is, understand what attributes it has (such as product gender, colour, length, style, material, fit, pattern and so on), place each product into the relevant area of the site and then promote more popular products over less popular ones, dependant on a user’s activity.

David Bishop, CTO & co-founder of LovetheSales, said: “The future of retail is highly connected with developments in Machine Learning.

“It’s an exciting time as it allows us to bring to market great services such as LovetheSales, which wouldn’t have been feasible 10 years ago – ultimately it’s consumers who will reap the benefits.”

Utilising Support Vector Machine technology, the system calls on over 100 million data points from training data that has been manually entered by the Lovethesales team over a 10 month period.

In less than eight hours Minerva reviews the entire product catalogue and tags each product individually with any relevant attributes.

She then places the items into the correct departments on the website, while promoting the stock she deems more likely to sell.

According to LovetheSales, to classify and promote 1 million products correctly, they would need to employ a team of 10 people for four years at a cost of around £1,250,000 just to complete this once. joins the ranks of Amazon, Facebook and Netflix as companies who have also invested heavily in Artificial Intelligence.