Onedot Platform News

The Onedot platform now supports a set of standard recipes for various common onboarding steps such as product classification, attribute extraction, attribute mapping, golden record determination, etc.

These recipes are part of the Onedot platform and can be easily added for custom data pipeline recipes like a code library, further increasing onboarding speed and ensuring improvements can be rolled out faster.

Improved calculation of confidence of suggestions in attribute mapping and product classification.

The reported confidence is now more intuitively understandable and describes the proposed target attributes and product classes more accurately in many cases.

For daily data pipelines, reports are now available that show statistics of product data delivered from different sources, such as number of SKUs per category, coverage between product master data and pricing data, etc.

This allows for easier traceability if there are problems with delivered data from content providers or other data suppliers.

With the help of various rules, it is now possible to configure exactly what should happen if an error occurs during onboarding.

With best-effort delivery, correctly processed suppliers or data sources are delivered, while erroneous ones are only reported, while with complete delivery, either all or nothing is delivered, which is useful when the product database is to be consistently populated with product master data and price data from a variety of sources.

The attribute mapping process of the Onedot platform has been further revised.

The learning process of the platform AI across supplier onboarding boundaries has been improved and for repeated onboardings only the differences to the last onboarding are shown.