What is the difference between the “Bulk file upload portal fees” and the “Classification fees”?

The Bulk file upload portal fees are the costs for each line of data in bulk CSV files uploaded to Zephyr through the bulk upload portal. Zephyr was designed for speed and volume over accuracy (originally for the purpose of “conveyor belt” processing of parcel customs clearance data - checking the accuracy and filling in the blanks with best matches).

The classification fees are for the AI classification engine we call Bluestone (which can be accessed through the classification API endpoint that is a part of Aura).

Where Zephyr is a fast matching algorithm, Bluestone is a true AI backed by Natural Language Processing and Machine Learning technologies, takes in a product description and returns up to 20 possible results for the user to select from or take the top match (this is the ’Auto Bluestone Classification).

When making an API call to Bluestone, there is an ‘Advanced Classification’ flag that can be turned on, by which our content team will go through the results from Bluestone ‘manually’, and do one of three things:

1. Core Classification - this means the content team will select the most appropriate matching commodity from Bluestone’s top 20 matches to return based on the product description
2. Enhanced Classification - in the case of more ‘complex’ commodities, the content team will do further research on the commodity before selecting the best match or adding the commodity to our system if it hasn’t seen it before.
3. Advanced Classification - for very complex or specific commodities that require expert classification, there is an hourly rate instead of a rate per commodity, as in these cases it can’t be predicted how much work is required to classify a commodity.

The costs for bulk file processing through the Zephyr portal would be similar to those of auto classification, but lower for higher volumes, where the costs for the classification API endpoint are fixed.