Category managers or cataloguing teams often receive catalogue data from multiple suppliers, each with different ways of naming the same part type/category. Consolidating and standardising this data is a manual, time-consuming and expensive process.

Part Type Classifier uses Partly’s catalogue of millions of parts and images to suggest the correct part type classification for each part.
‍

Benefits

By standardising part type classifications, category managers can:

1. Manage part information more effectively by adding part-type specific attributes and updating part information in bulk per part types
2. Provide a cohesive category tree in their catalogue and their e-commerce website.
3. Utilise advanced data enrichment techniques
   ‍

How it works

1. Category managers import their data into Partly PIM via CSV files, marketplace integration, or a 3rd party inventory manager.
2. Machine-learning algorithms will suggest a part type for their parts based on the unstructured part type/category imported, or the titles and Images of the linked products
3. Users either confirm if the prediction is correct or can submit a correction.
4. The machine-learning improves based on user input.
   ‍

FAQs

How are Partly part types classified?‍

Partly has a vast catalog comprising millions of meticulously researched parts, each carefully assigned with a specific part type. Within the Partly taxonomy standard, there are approximately 28,000 distinct part types. Leveraging this extensive dataset, Partly's Part Type Classifier employs artificial intelligence to accurately predict the Part Type utilising an image or textual input.
‍

How does the system work?

1. Simply load in a list of part numbers with text descriptions and/or images.
2. Partly’s Part Type Classifier AI engine will generate the predicted part type for every part number.
3. For each part, confirm whether the predicted part type is correct, or select from other suggested part types.

‍