Product Description from Feed

Input Variables Structure

The workflow is built around a series of specific input variables that guide the AI in generating appropriate descriptions:

1. ProductName: The complete name of the product as it appears in your catalog.
2. ProductType: The category or type of product, which helps contextualize the description.
3. ProductBrand: The manufacturer or brand name associated with the product.
4. ProductAttributes: A comprehensive list of product specifications, features, materials, dimensions, and other relevant details that should be highlighted in the description.
5. SpecificInstructions: Any particular guidance or requirements for the product description, such as tone preferences, emphasis on certain benefits, or marketing angles to explore.
6. OutputLanguage: The desired language for the final product description, allowing for multilingual catalog support.

The Processing Flow

The workflow consists of two primary nodes that work together to process your product data:

Local Upload Node

This component serves as the initial data entry point, allowing users to upload product information from local sources. While marked as deprecated in the current version and scheduled for removal, it currently functions as the initial step in the workflow process.

Product Description Generation Node

This is the core processing component where the magic happens. When activated, this node:

1. Receives and processes all input variables from either the Local Upload node or direct user input.
2. Applies sophisticated natural language processing using GPT-4o, configured with specific parameters to optimize output quality:

• Carefully calibrated temperature setting to balance creativity and factual accuracy
• Token limits to ensure descriptions are comprehensive yet concise
• Response formatting guidelines that create consistent, scannable description

Follows an internal guideline-based approach to ensure descriptions:

• Focus on customer benefits rather than just listing features
• Highlight how products transform or improve the customer’s life
• Maintain brand voice consistency
• Include all relevant product attributes in a natural, engaging way
• Adhere to e-commerce best practices for conversion optimization

Produces a finalized product description through the “ProductDescription” output variable, ready for implementation in your e-commerce platform.

Data Flow and Processing Logic

The workflow connects these nodes through a defined edge that establishes the path data follows from input to output. When you submit product information, it flows from the initial input stage through the sophisticated language processing engine, which transforms raw data into compelling copy.

The system uses contextual understanding to prioritize the most important product features and present them in ways most likely to resonate with potential buyers. Rather than simply listing specifications, the workflow constructs narratives around product benefits, addressing customer pain points and desires.