We can use AI Assistant to add new transformations and aggregations to an existing pipeline. However, we can't upload the existing table and use AI Assistant to update the schema.
Example scenario: I upload a pipeline from the gallery and I want to modify the schema of one of the tables, for example I want to remove one column from the 'user' table. Because I cannot upload existing tables to AI Assistant and ask him to update the schema, I need to re-generate all the tables for the scenario, which might take a while and may or may not include the 'user' table. If the 'user' table is generated, then I can ask AI Assistant to update the schema, regenerate the data, and add that to the pipeline.