So, I have an ERP system and a whole ecosystem of other systems around it. How is supposed to help me a set of data products to address the data quality issues? Do I need to create a series of data products (e.g. to replace a data warehouse or data lake) to address the issues?

Moreover, data quality is judged in respect to a set of needs that can vary across systems. How do I address this?

From what I understand, a data product will make sure that the data it receives is consisted with its internal expectations. What happens if suddenly new data with different rules need to be consumed? I suppose the values will be rejected. How can this address the quality existing in source systems?

Can you estimate how much logic needs to be included in a product to process the data via its pipelines, to validate the data, to cope with exceptions or to implement the changes?

My general feeling is that a data product will become a little monster that attempts to handle too much than to be pragmatic. In the end I need the data as reflected in the source system, and they should be clean.