Automated Product Description Generator

About

The client is an innovative SaaS-based Product Information Management (PIM) and Digital Asset Management (DAM) platform. Based in Manchester, UK, the client helps businesses to expand their online sales rapidly through managing their product information more efficiently. Having built the platform starting in 2016, the client was looking for an expert partner to scale the solution to integrate with newer technologies such as AI. We provided the expertise towards the same.

Challenge

While the platform had features for manually entering product descriptions, the client wanted us to build an AI-powered workflow for generating product descriptions from product metadata. The specific challenges addressed were:

  • Tedious and time-consuming process of creating product descriptions manually
  • Limited resources to create product descriptions for a large product library
  • Inconsistent product descriptions across the platform
  • Inability to keep up with SEO best practices and optimization due to the volume of products

Solution

Our team of AI and machine learning experts enabled the modernization of the platform. Features we built include:

  • AI-powered workflow that uses product metadata, images, and OpenAI to generate and update SEO-optimized product descriptions for the entire product library in the brand’s voice
  • Using Natural Language Generation (NLG) models to automatically generate product descriptions from structured metadata such as product name, category, features, and benefits
  • Training the NLG models with brand-specific writing styles, tone, and language to ensure consistency across all product descriptions, for multiple product types such as apparel, furniture, accessories, etc
  • Incorporating image analysis and recognition to provide more accurate and descriptive product descriptions
  • Implementing a data pipeline to extract, transform, and load product metadata into the workflow and push updated product descriptions to the eCommerce platform of customers.

 

The metadata associated with each product (e.g., product name, brand, features, specifications, etc.) was used to provide the AI model with the necessary information to generate the product description. The metadata can be fed into the model as input features and used to guide the generation of relevant and accurate product descriptions.

 

Product images were also used to provide additional information to the AI model. By analyzing the product images, the AI model can identify key features and attributes of the product and incorporate them into the product description.

 

We leveraged OpenAI – a powerful AI platform that can be used to generate high-quality natural language text. By integrating OpenAI’s language generation capabilities into the product description workflow, it is possible to generate product descriptions that are both accurate and engaging.

 

The product descriptions generated by the AI model were optimized for search engine optimization (SEO) by incorporating relevant keywords and phrases into the text. This can help improve the visibility of the product in search results and drive more traffic to the eCommerce platforms of customers. The solution automatically analyzed output for readability, originality, and SEO-friendliness.

Super charging the platform with AI capabilities provided several benefits such as:

  • Significant time savings and reduced workload for content teams
  • Consistent and high-quality product descriptions across the platform
  • Improved SEO optimization and increased organic traffic from search engines
  • Ability to quickly and easily update product descriptions as product details change
  • Enhanced customer experience with more informative and engaging product descriptions

Webinar ON-DEMAND

Customizing Large Language Models & Generative AI for Business