Visual AI for Retail

Image is everything

It takes the average person around 25 minutes to create product information for a product. If they were to add 10k products to a PIM it would take 20 months! Our visual AI can do this in a fraction of the time, more accurately, with greater consistency at far less cost - welcome to a new veritical in the sector.

Contact UsRequest Demo

What problem does PrePIM AI solve?

The idea for PrePIM AI came up when we visited on the largest pureplayers in France, Showroom Prive, who are employing an army of people during peak season to write content in multiple languages for the fashion they sell online.

This content writing process is time consuming, expensive and inconsistent due to the interpretation of each writer. What is a mini dress to you might not be to someone else! We've identified that whether you are a small retailer without a dedicated full time product information classification person, or a large retailer with a big team, time getting product onto the site is money. Errors or lack of detail in the product titles and descriptions, are lost opportunities and stock left on in the warehouse shelves. PrePIM AI generates a wide range of attributes and features that can be combined to make feature rich titles and descriptions. It does so, quickly, accurately and consistently. Because it can be trained by people, it will learn the best attributes to apply to your stock when new trends come along. It's a system with a thirst for knowledge.

How it works

It is very simple, you upload your pictures in our interface and PrePIM AI will do the hard work by "seeing" those images and creating feature rich attributes; from colours, to sleeve length, to neck type, etc. It can also draft titles and descriptions according to your editorial preferences. The user interface is currently available in English and French, the platform is capable of working in any language, so it's possible to translate your product information for sale in any country. Pre-PIM AI currently works with Men's and Women's fashion, accessories and shoes.

The Benefits

For large retailers You are able to capture your human capital and get your best people to work with Pre-PIM AI to train it to the very highest standards. Giving you fast, accurate, feature rich and consistent attributes, titles and descriptions for your Products. API integration to your PIM makes the process seamless. Every time new products or images are added, Pre-PIM AI can add the necessary attributes straight back into your PIM once it's been approved by a human

For smaller retailersYou might not have enough products being added to have a dedicated human resource for creating product information. The downside to this is that if multiple people are involved in creating the data, i.e. whomever is free to do it, the continuity of attributes, accuracy and speed of adding product information is impacted. Using Pre-PIM AI to create the data means you've got a reliable, accurate resource on hand to help 24/7. If you don't have language capability in-house you can use Pre-PIM to create the data in multiple languages. Export the file and import it straight to your website, if you don't have a PIM to import it to.

Our Visual AI humble beginning

Our visual AI was born a few months ago out of a will to help fashion retailers audit their feeds. The reason behind that is we discovered time and time again that feeds were littered with mistakes. One feed we encountered had 15,000 products wrongly labelled as t-shirts! Those mistakes are costly for your PPC advertising as if a product is mislabelled in the feed it will show for irrelevant searches.

So Dream Agility Intelligent Vision (DAIV) is a visual AI feed auditor! It is able to look at an image and compare it to the data and decide whether the data is correct in terms of colour, Google product category etc. It is destined for marketers to make sure their product feed is as accurate as possible in order to make the most of their shopping campaigns.

PrePIM AI is the antithesis of DAIV. Instead of validating data it creates it from scratch, and it is aimed at merchandisers.

Contact UsRequest Demo