The architecture principle is to avoid unnecessary data movement when the external platform can be queried or shared securely. Query federation with acceleration (Caching) aligns with the zero-copy model because Data 360 can expose or query governed data without building another extract pipeline. That is important when teams want freshness, reduced duplication, and lower operational burden while still respecting permissions and platform boundaries. The distractors fall short because they either move the problem into the wrong system, add needless duplication, ignore Data 360 object relationships, or rely on a feature built for a different lifecycle stage. In a real implementation, those choices usually create brittle pipelines, stale data, security exposure, or segments that look correct on paper but fail when activated. Thinking like an architect, the selected option places the logic where Data 360 can govern it and reuse it reliably. This is the nuance exam questions often test: the platform capability must match both the technical layer and the business timing requirement, not just sound related to data.

The segmentation and activation design starts with grain: who or what the audience represents, and which attributes must travel with it. Flow works because Data 360 segmentation and activation must respect audience grain, relationship paths, and activation payload rules. A segment can qualify the audience, but activation determines which related attributes or contact points are actually sent downstream. The distractors fall short because they either move the problem into the wrong system, add needless duplication, ignore Data 360 object relationships, or rely on a feature built for a different lifecycle stage. In a real implementation, those choices usually create brittle pipelines, stale data, security exposure, or segments that look correct on paper but fail when activated. Thinking like an architect, the selected option places the logic where Data 360 can govern it and reuse it reliably. This is the nuance exam questions often test: the platform capability must match both the technical layer and the business timing requirement, not just sound related to data.

Harmonization is the Data 360 modeling process that maps differing source schemas into a consistent, standardized data model. Salesforce explains that source schemas preserved during ingestion are harmonized during data modeling by mapping Data Lake Objects to Data Model Objects in the Customer 360 Data Model. This enables semantically equivalent source attributes—such as user, subscriber, or contact representations—to be interpreted consistently downstream. Transformation is concerned with changing or deriving data values or structures, while segmentation identifies audiences based on criteria. Unification occurs later through Identity Resolution, where records representing the same person or account are linked into unified profiles. Because the problem is inconsistent schema terminology rather than duplicate identities, harmonize precisely describes the required process.

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

The segmentation and activation design starts with grain: who or what the audience represents, and which attributes must travel with it. Activation membership to select the Loyalty Program Member DMO and its associated fields works because Data 360 segmentation and activation must respect audience grain, relationship paths, and activation payload rules. A segment can qualify the audience, but activation determines which related attributes or contact points are actually sent downstream. The distractors fall short because they either move the problem into the wrong system, add needless duplication, ignore Data 360 object relationships, or rely on a feature built for a different lifecycle stage. In a real implementation, those choices usually create brittle pipelines, stale data, security exposure, or segments that look correct on paper but fail when activated. Thinking like an architect, the selected option places the logic where Data 360 can govern it and reuse it reliably. This is the nuance exam questions often test: the platform capability must match both the technical layer and the business timing requirement, not just sound related to data.

The modeling decision should make the data understandable, reusable, and correctly related before segmentation or activation depends on it. Map standard fields to the Sales Order DMO and add a custom attribute for the fee. fits because Data 360 depends on mappings and relationships between data lake objects and data model objects. Correct modeling lets the same attribute mean the same thing across source systems and prevents downstream users from building logic on ambiguous fields. The distractors fall short because they either move the problem into the wrong system, add needless duplication, ignore Data 360 object relationships, or rely on a feature built for a different lifecycle stage. In a real implementation, those choices usually create brittle pipelines, stale data, security exposure, or segments that look correct on paper but fail when activated. Thinking like an architect, the selected option places the logic where Data 360 can govern it and reuse it reliably.

The lifecycle point is that Data 360 configuration is metadata, but it still has product-specific packaging and activation steps. Install the data kit from the Data 360 Setup menu to map and activate the components. is the right fit because Data 360 components are not fully usable just because generic deployment succeeded. The consultant must complete the Data 360-specific lifecycle step so streams, mappings, insights, and rules are correctly registered in the target environment. The distractors fall short because they either move the problem into the wrong system, add needless duplication, ignore Data 360 object relationships, or rely on a feature built for a different lifecycle stage. In a real implementation, those choices usually create brittle pipelines, stale data, security exposure, or segments that look correct on paper but fail when activated. Thinking like an architect, the selected option places the logic where Data 360 can govern it and reuse it reliably.

The design point is to preserve source fidelity while shaping data only where Data 360 processing needs it. Here, B2C Commerce Starter Bundles plus a custom extract fits because it changes the shape, keying, or refresh behavior at the Data 360 layer instead of forcing the source system to carry an analytics-specific design. In production, this keeps the upstream application simpler and gives the data team a repeatable way to prepare records for mapping, identity resolution, insights, or segmentation. The distractors fall short because they either move the problem into the wrong system, add needless duplication, ignore Data 360 object relationships, or rely on a feature built for a different lifecycle stage. In a real implementation, those choices usually create brittle pipelines, stale data, security exposure, or segments that look correct on paper but fail when activated. Thinking like an architect, the selected option places the logic where Data 360 can govern it and reuse it reliably.

The segmentation and activation design starts with grain: who or what the audience represents, and which attributes must travel with it. Create a reusable container block with common criteria. works because Data 360 segmentation and activation must respect audience grain, relationship paths, and activation payload rules. A segment can qualify the audience, but activation determines which related attributes or contact points are actually sent downstream. The distractors fall short because they either move the problem into the wrong system, add needless duplication, ignore Data 360 object relationships, or rely on a feature built for a different lifecycle stage. In a real implementation, those choices usually create brittle pipelines, stale data, security exposure, or segments that look correct on paper but fail when activated. Thinking like an architect, the selected option places the logic where Data 360 can govern it and reuse it reliably. This is the nuance exam questions often test: the platform capability must match both the technical layer and the business timing requirement, not just sound related to data.

The design point is to preserve source fidelity while shaping data only where Data 360 processing needs it. Here, When a column is added or removed, the CRM Connector performs a full refresh. fits because it changes the shape, keying, or refresh behavior at the Data 360 layer instead of forcing the source system to carry an analytics-specific design. In production, this keeps the upstream application simpler and gives the data team a repeatable way to prepare records for mapping, identity resolution, insights, or segmentation. The distractors fall short because they either move the problem into the wrong system, add needless duplication, ignore Data 360 object relationships, or rely on a feature built for a different lifecycle stage. In a real implementation, those choices usually create brittle pipelines, stale data, security exposure, or segments that look correct on paper but fail when activated. Thinking like an architect, the selected option places the logic where Data 360 can govern it and reuse it reliably.

Once a predictive model is activated in AI Models, formerly Einstein Studio, Salesforce provides supported inference channels for operational use. Predict jobs can run in batch or streaming mode and write prediction outputs to a Data Model Object, where downstream Salesforce processes can consume them. Salesforce also surfaces activated predictive models as actions in Flow Builder so predictions can drive decisions and automation directly in business workflows. That combination is exactly what option D describes. Manually exporting and re-importing CSV files adds latency and unnecessary data movement. Identity Resolution is for linking and reconciling profile records, not for invoking predictive models. Embedding an AI Models URL in a Lightning component likewise does not constitute a supported prediction-inference pattern. Therefore, predict jobs or Flow Builder are the standard operationalization methods.

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