Amber Engine Insights & AI

“Product Insights” give a brand valuable information about why a product is selling or not selling, and how to improve it’s sales potential. We had partnered with a company to build an external platform that would allow our customers to feed it product data and roughly achieve this, but it was always within our roadmap to fold this functionality directly into our PIM and allow our users to have an integrated insights flow. At the end of 2022 I designed a proof-of-concept of this integration in the screens below.

The Master Insights Section of the product

  • Provides an overview of Insights activity
  • A dashboard experience with various metrics and reporting
  • Tabs for settings, attribute mapping, and perhaps Insights Engine assignment


Individual Product Insights

The system could be set to recognize the types of subjective attributes that can affect sales, and determine scores for each based on competitive products. Each attribute can be weighted, based on how much it affects sales potential. Product Title (name) might be given 4 point, while feature bullets is only worth 1 or 2 points. Product Description, Primary Image, and Additional Images all have theor own weight that add up to a possible 10 points. A ten-out-often score would mean it is the best performing product in it’s category. The insight’s engine would measure how well each attribute performs against it’s contemporaries. In this example, this product’s Title scores 1.5 out of a possible 3.99.

The basic exercise is to try to make the donut as green as possible!

The user would achieve this by clicking on each component attribute and editing it. The insights engine would give the user insight as to why it’s not performing well and how to fix it. The process is to use that information to enhance each given attribute, then refresh the score and see if you’ve made things better. In the slideshow at the bottom of this page, we have an example of using ChatGPT and other AI bots to make this enhancement automatic!

Insights within the Product Page (PDP). A new tab within the PDP that houses insights across several attributes for that specific product.

Attribute Insights

Imagine turning on insights at the column (attribute) level. Not for every attribute of course. For example, insights on dimensional data would be pretty useless… a product is as large or small as it needs to be. But for “soft” or subjective data, like product name, product descriptions, bullets, imagery, etc, insights can be a powerful tool.

After enabling insights for a specific attribute, the user will be alerted to pending insights for a given field with a small widget. Clicking the widget opens a modal experience that allows the user to enhance the data in that field based on calculated insights.

Using AI to improve Product Sales within the PIM