Study guide
12.3 Applying Predictive & Generative AI Tooling to Customer Scenarios
Salesforce Certified Data 360 Consultant formerly Data Cloud Consultant
Track your progress as you read
Finished sections get a check mark.
1 Chapter 1: Solution Overview & Enterprise Value
1.1 Enterprise Customer 360 & Data Cloud Business Value 1.2 Architectural Foundation: Zero-Copy, Hyperforce, Lakehouse & Data Pipeline 1.3 Platform Terminology, Core Entities & Data Lifecycle Stages 1.4 Data 360's Foundational Role in Predictive & Generative AI
2 Chapter 2: Ethics, Privacy & Governance Fundamentals
2.1 Regulatory Compliance, Data Privacy & Ethics in Customer Data 2.2 Consent Management & Preference Enforcement in Data Cloud 2.3 Data Retention, Deletion, and Privacy Right-to-be-Forgotten (RTBF)
3 Chapter 3: Security, Permissions & Tenant Architecture
3.1 Standard Permission Sets, User Roles & Functional Access 3.2 Data Spaces Architecture, Resource Allocation & Tenant Isolation
4 Chapter 4: Administration, Monitoring & Lifecycle Management
4.1 Data Kits & Package Automation for Cross-Org Deployment 4.2 Auditing, Status Monitoring, Error Logs & Data Explorer 4.3 Sandboxes, Change Management & Metadata Deployment
5 Chapter 5: Ingestion Connectors & Source Integration
5.1 Standard Bundles & Salesforce CRM Connectors (Sales/Service/Marketing) 5.2 Cloud Storage (S3, GCS, Azure Blob) & Streaming Ingestion (Web/Mobile SDK, Ingestion API) 5.3 Zero-Copy Data Federation & Snowflake/BigQuery/Databricks Sharing
6 Chapter 6: Data Lake Objects (DLO) & Data Transformations
6.1 DLO Configuration, Primary Keys, Event Time & Data Stream Categories (Profile, Engagement, Other) 6.2 Streaming & Batch Data Transforms: Formula Fields, SQL Transforms & Data Prep Recipes
7 Chapter 7: Data Model Objects (DMO) & Canonical Harmonization
7.1 Customer 360 Canonical Data Model & Subject Areas 7.2 DLO-to-DMO Mapping Patterns, Cardinality & Custom DMO Design 7.3 Data Validation, Schema Maintenance & Field Mapping Best Practices
8 Chapter 8: Identity Resolution Architecture & Match Rules
8.1 Identity Resolution Concepts, Consolidation Rate & Graph Modeling 8.2 Match Rules & Match Criteria: Exact Match vs. Fuzzy Matching (Deterministic & Probabilistic) 8.3 Multi-Party Identification, Anonymous-to-Known Transitions & Contact Point Matching
9 Chapter 9: Reconciliation Rules & Unified Profile Governance
9.1 Reconciliation Rules: Most Frequent, Source Priority & Last Updated 9.2 Unified Profile Schema (Individual & Contact Points) & Resolution Troubleshooting
10 Chapter 10: Segment Architecture & Audience Design
10.1 Segmentation Canvas, Segment Types (Standard UI & Flow) & Query Processing 10.2 Direct vs. Related Attributes, Cardinality & 1:N Filtering Traps 10.3 Segment Schedule, Refresh Frequencies & Exclusion Logic 10.4 Segment Membership Analysis, Population Drift & Sizing Validation
11 Chapter 11: Calculated & Streaming Insights
11.1 Calculated Insights (CI) Architecture, SQL/Builder Syntax & Dimensions vs. Metrics 11.2 Streaming Insights (SI) Architecture, Windowing (Tumbling/Sliding) & Real-Time Signals 11.3 CI vs. SI Architectural Decision Matrix & Performance Optimization
12 Chapter 12: Data Enhancements, Sharing & Analysis
12.1 Referencing Data 360 Data from Other Systems: Enrichments, Related Lists, Data Graphs & APIs 12.2 Reports, Dashboards & Tableau Analysis on Data 360 Data 12.3 Applying Predictive & Generative AI Tooling to Customer Scenarios
13 Chapter 13: Activation Targets & Outbound Activation
13.1 Activation Targets: Marketing Cloud, External Platforms, Cloud Storage & Data Spaces 13.2 Activation Configuration: Audience Mapping, Related Attributes & Payload Optimization 13.3 Activation Schedules, Publishing Cadence & Incremental vs. Full Refresh 13.4 Activation Troubleshooting: Error Logs, Rejection Reasons & Delivery Failures
14 Chapter 14: Real-Time Data Actions & Flow Automation
14.1 Data Action Targets & Webhook/Platform Event Architecture 14.2 Data Cloud-Triggered Flows & Real-Time Business Orchestration 14.3 End-to-End Enterprise Scenario Walkthrough: Ingestion to Closed-Loop Action
12.3 Applying Predictive & Generative AI Tooling to Customer Scenarios
Key Takeaways Explain This Topic AI
AI Models, formerly Einstein Studio, is the Data 360 workspace for predictive models, bring your own model connections, bring your own large language models, and retrievers.
Model Builder supports three paths: build a predictive model from scratch with clicks, connect an external predictive model from Amazon SageMaker, Google Cloud Vertex AI, or Databricks, and connect an external foundation model.
A predictive score returned into Data 360 becomes ordinary data: filterable in the Segmentation canvas, referenceable in calculated insights, and consumable by Flow and prompt templates.
Retrievers created in the AI Models tab query vector, hybrid, or enriched search indexes and are what supply unstructured grounding to prompt templates and Agentforce agents.
Predictions are personal data: consent captured through the Consent API, data space scoping, and the Einstein Trust Layer all continue to apply to model inputs and outputs.
Last updated: September 2026
12.3 Applying Predictive & Generative AI Tooling to Customer Scenarios
Quick Answer: Section 1.4 covered why Data 360 underpins AI. This section covers what you configure . AI Models — the workspace formerly called Einstein Studio — is where a consultant creates predictive models from scratch, connects external models with bring your own model (BYOM) , connects external foundation models with BYOLLM , and builds retrievers over search indexes. Predictions come back into Data 360 as data, which makes them usable in segmentation , calculated insights , Flow , and prompt templates . The exam tests scenario matching: given a business requirement, which tool, and what has to be true about the data first.
This is the "Apply predictive and generative AI tooling capabilities of Data 360 to customer scenarios" objective inside Data Enhancements, Sharing, and Analysis (18%) .
AI Models: The Workspace
AI Models is the single place inside Data 360 for model management, regardless of model type or where the model runs. What a consultant configures there:
Capability What It Is Typical Scenario
Predictive model from scratch Clicks-not-code model trained on a Data 360 object Churn likelihood, propensity to buy, expected order value
Bring your own model (BYOM) A predictive model trained on Amazon SageMaker, Google Cloud Vertex AI, or Databricks, connected to Data 360 The data science team already owns a validated model and its MLOps pipeline
Bring your own large language model (BYOLLM) An externally hosted foundation model connected for generative use The organisation has a contracted or fine-tuned LLM it must use
Retrievers Query components over vector, hybrid, or enriched search indexes Grounding an agent or prompt template on knowledge articles, manuals, or transcripts
The zero-copy principle carries through to BYOM: the external platform trains against Data 360 data without ETL , and once deployed, predictions update continuously and land back in Data 360.
- The Predictive Workflow End to End
- Step 1 — Unify First
A predictive model trained on Individual__dlm learns from duplicates. Train on the unified profile. This is the single most consequential prerequisite, and it is why identity resolution (Chapters 8 and 9) precedes AI in every sensible implementation sequence.
Step 2 — Engineer Features Inside Data 360
The documented pattern is to flatten behaviour and history into a one-row-per-customer feature set using calculated insights and data transforms : spend in the last 90 days, order count in the last 12 months, sessions in the last 30 days, days since last case, average resolution time. Keeping this work inside Data 360 avoids external joins, keeps training fast and scalable, and keeps the data under in-platform governance.
Step 3 — Train
Either build from scratch in AI Models — select the source object, select the outcome field, let Data 360 train and score — or connect the external model with BYOM.
Step 4 — Operationalize the Prediction
This is the step scenarios actually test. Once a score is in Data 360, it is just an attribute:
Segmentation. Filter the canvas on Churn Propensity > 0.7 to build a retention audience, then activate it to Marketing Cloud Engagement.
Calculated insights. Reference the score in a further aggregation — average churn risk by region, revenue at risk by loyalty tier.
Flow. A Data 360-triggered flow reads the score and creates a proactive task for the account executive.
Prompt templates and agents. A grounded prompt can reference the score so an agent's suggested next step reflects the customer's risk.
Enrichment. Copy the score onto the Account or Contact so ordinary CRM reports, list views, and validation rules can use it.
Scoring runs in batch or streaming mode; streaming scores records as data changes, which matters when the decision is being made during a live interaction.
- The Generative Workflow End to End
- Structured Grounding
For attributes that already live in the model, grounding is declarative: ground the prompt template on DMO merge fields using the two supported relationships — copy fields and related lists of DMOs on Salesforce objects (section 12.1).
Unstructured Grounding
For content that is not tabular — knowledge articles, contracts, manuals, call transcripts, and also audio, video, and images — the pipeline is:
Ingest the unstructured content into Data 360.
Map it to standard DMOs or unstructured data model objects (UDMOs) .
Chunk it into meaningful passages.
Embed the chunks as vectors.
Index them as a vector , hybrid , or enriched search index.
Create a retriever in the AI Models tab over that index.
Reference the retriever from a prompt template or an Agentforce agent for RAG.
Choosing the index type is a real consultant decision. Pure semantic recall over conceptual questions suits a vector index. Content full of exact tokens — SKUs, part numbers, policy codes — suits a hybrid index, because pure semantic matching will drift off an exact identifier. Where recall quality is paramount, an enriched index generates three chunks per passage: the original text, a metadata chunk, and a chunk of questions the passage can answer.
Scenario-to-Tool Matching
Business Requirement Correct Tool Why Not the Alternatives
Score every customer's likelihood to churn next quarter Predictive model in AI Models (scratch or BYOM) A calculated insight computes history, not likelihood
Reuse a validated model the data science team built on SageMaker BYOM Rebuilding from scratch discards validated work and its MLOps pipeline
Build a retention audience from the churn score Segmentation canvas filtered on the score attribute The score is already data; no further modelling needed
Let a service agent ask questions of 40,000 knowledge articles Search index + retriever + prompt template Merge fields cover structured attributes, not free text
Surface a lifetime value score on the Account page Copy field enrichment A retriever is for unstructured content
Give an agent the full customer context in under a second Data graph An insight is batch; a retriever serves unstructured content
Notify an account executive when risk crosses a threshold Data 360-triggered flow reading the score An activation delivers audiences, not single-record alerts
Governance Obligations That Survive the Move to AI
A consultant who presents AI tooling without the governance frame has given an incomplete answer, and the exam rewards the complete one:
Consent. A prediction about a person is personal data. A profile that opted out of processing through the Consent API processing action should not be feeding a model or an agent response, and a profile deleted with shouldforget must not survive as a model input.
Data spaces. Models, retrievers, and search indexes are scoped. A brand's model does not silently read another brand's data.
The Einstein Trust Layer sits between the grounded prompt and the model, applying protections such as masking and zero data retention with the model provider. It requires Data 360 enabled.
Explainability and drift. A score that was accurate at launch degrades as behaviour shifts. Schedule retraining and monitor the score distribution the same way you monitor consolidation rate.
Exam Traps & Consultant Pitfalls
Trap 1: Modelling on fragmented data. Training on source-level individuals rather than unified individuals reproduces duplicates as features. Unify first, always.
Trap 2: Treating a calculated insight as a prediction. A calculated insight summarizes what happened. A model estimates what will happen. "Customers who have not purchased in 90 days" is an insight; "customers likely to churn in 90 days" is a model.
Trap 3: Using a retriever for structured attributes. If the value is a field on a DMO, ground with merge fields. Retrievers exist for chunked, embedded, unstructured content.
Trap 4: Forgetting where the score lives. A prediction is written back into Data 360 and is then usable in segmentation, insights, flows, and enrichments. Answers that propose exporting scores to a file and re-ingesting them are recreating a pipeline the platform already removed.
Trap 5: Choosing a vector index for exact-token content. Part numbers and policy codes need lexical matching too — that is what the hybrid index is for.