The correct approach is to use the Tableau Semantics connector from Tableau Cloud. Salesforce explicitly supports connecting Tableau Desktop and Tableau Cloud directly to semantic models defined in Data 360/Tableau Semantics. This allows downstream Tableau authors to consume the same governed measures, dimensions, calculations, relationships, metrics, and business definitions that Tableau Next uses.

Option B would undermine the principal benefit of the semantic layer. Reconnecting directly to Data 360 objects and recreating joins would force analysts to reconstruct business logic and could lead to inconsistent calculations or relationships across analytical applications. Option C is also incorrect because Data Kits are Salesforce/Data 360 deployment mechanisms; they are not used to install semantic models into Tableau Cloud as independent Tableau data models.

This interoperability capability is strategically important: Tableau Semantics provides one governed analytical definition layer that can be consumed across Tableau Next, Tableau Cloud, Tableau Desktop, AI experiences, and other supported Salesforce applications. Analysts can therefore create custom Tableau Cloud visualizations without duplicating semantic logic.

Reference/Topics: Embedding, Cross-Cloud, and Interoperability -> Tableau Semantics Connector -> Tableau Cloud -> Reusable Semantic Models.

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