Data Pro—now increasingly referred to under Semantic Modeling functionality—depends on relationships already defined between the data model objects participating in a generated calculated field. Salesforce explicitly states that a predefined relationship must exist between DMOs for Data Pro to create a calculated field spanning those objects. The agent cannot construct a valid cross-object calculation when the required relationship does not exist.

Therefore, if calculations involving Sales and Customer work but introducing Territory causes the operation to fail, the most likely issue is that Territory is not relationally connected to the relevant objects.

Good metadata and field descriptions improve semantic clarity and AI accuracy, but insufficient descriptions are not the hard technical prerequisite being tested. Likewise, Salesforce does not document a general two-DMO maximum corresponding to option C.

The broader architectural principle is that generative semantic-model functionality does not replace proper data modeling. AI features consume the semantic model's existing object structure, metadata, and relationships. Consultants must establish valid relationships before expecting Tableau Agent to synthesize calculations involving fields from multiple objects.

References/Topics: Data Setup - > Semantic Models - > DMO Relationships - > AI-Generated Calculated Fields - > Data Pro/Semantic Modeling.

===========================================================================