PIM, AI and Machine Learning, some opportunities

Machine Learning and Artificial Intelligence, or AI, are topics that are booming right now. Can you also do something with AI and Machine Learning in the area of product management? And if so, how can you make it work for you in conjunction with your PIM system? 

At ConnectingTheDots, we have been working on AI and Machine Learning in combination with product management for some time. In this blog, we want to present some practical possibilities that are relatively easy to apply in a PIM system. 

Image recognition

You can apply Machine Learning to product images. Based on a product image, Machine Learning can be used to obtain interesting information about a product. Think of color information, determining the product type and several useful product attributes. This information can then be used for a number of things. 

First, product classification. Based on the information on a product image, such as product type,  a product can be classified in a PIM solution. Which can be an efficiency improvement if employees have to do this manually in PIM. A service like Amazon Rekognition, for example, is able to provide this info.   

Secondly, you can use the product attributes generated through Machine Learning. These can be used to automatically fill in missing product attributes for products in a PIM solution. 

Product content populating

AI can be used to enrich product content. This includes product descriptions, product titles and so on. Here is how it works. Product attributes are sent to an AI tool such as ChatGPT from PIM. Think of attributes such as product type, brand, color, material, info on season etc. Based on this, an AI tool generates a product description that can be used on the product. Besides an efficiency advantage, (after all the text does not have to be written by a product content specialist) is the fact that unique product content is generated. 

Of course, you can also go one step further. For example, you can also use an AI tool to have multiple descriptions generated for a product. For example, to generate personalized product descriptions.  Of course, you could combine both of the above examples. So first generate product attributes based on a product image. And then based on these obtained attributes you can have product descriptions generated by an AI tool.

In short, Machine Learning and AI are powerful tools in the domain of product management. The choice is yours! However, remember one thing. Make sure you always (randomly) check the output. AI and Machine Learning are not perfect (yet).