Build the Foundation with Data 360

Learning Objectives

After completing this unit, you’ll be able to:

Explain the role of Data 360 as the foundational data layer for agentic analytics.
Describe the preparation and setup phases of a Data 360 implementation.
Differentiate between the preparation, harmonization, and unification phases of data structuring.

The Importance of a Strong Foundation

Data 360 (formerly known as Data Cloud) is the heart of the Salesforce ecosystem. It’s a powerful data engine that seamlessly expands to process vast amounts of data in real time to ensure that insights stay up to date no matter how large or complex your data becomes.

Imagine trying to teach an AI agent about your company’s performance using five different spreadsheets, a SQL database, and your customer relationship management (CRM) software. Without Data 360, the agent sees these data sources as disconnected silos. Data 360 brings this data into an environment where it can be grounded. AI grounding connects AI models to real-world, verifiable data to ensure accuracy and relevance, and prevent the inaccurate outputs that are common in generic AI models.

Prepare the Foundation

Before you click a single button in Data 360, you must define your data strategy and create a roadmap for your agentic analytics implementation. A successful agentic analytics framework isn’t just about connecting systems. You need to ensure that data is governed, accessible, and aligned with your business goals from the start.

Plan Your Data Strategy

A great data strategy acts as the blueprint for your grounding layer. To guide your strategy, focus on these key pillars.

Business alignment: Get clear about what your users need to accomplish. They might want to reduce customer churn by triggering automated flows, empower sales reps with real-time health scores, or do both simultaneously to drive growth.
Data governance and security: Establish clear data ownership and define how your information should be protected. Data governance provides the framework to keep your data safe and organized, allowing you to implement a robust content management strategy for regulatory compliance. Use these strict access controls to maintain data privacy across business units and determine if you need specific governance policies.
Inventory and quality: Audit your data sources. Identify the source of truth for critical fields like email addresses or customer IDs to ensure the quality and accuracy of your downstream insights.
Organizational readiness: Ensure you have the right stakeholders involved, from IT engineers who understand the source schemas to business leaders who understand the desired outcomes.

Illustrated clipboard with a checklist representing the key pillars of a Data 360 data strategy.

To follow a checklist, see Data 360 Checklist and Considerations.

Build the Foundation of Truth

With your strategy in place, the next step is to build the data structure that supports your agentic analytics experience and ensures your agents have a reliable source of truth to draw from.

Because data and business needs are constantly evolving, building this foundation is an iterative process. Begin in a sandbox environment, where you can safely prototype your data connections and logic. This sandbox-first approach allows you to set up your initial structure, test how it handles real-world scenarios, and fine-tune the details.

While the Data 360 Implementation Guide details the complete technical steps of establishing the grounding layer, let’s take a look at each of these phases to understand why they’re vital for building a data foundation that’s truly AI-ready.

Set Up Your Secure Data Environment

To get started, you must first activate your core platform capabilities, assign secure team roles, and establish data boundaries. These steps prepare your environment for the configuration stages that follow.

Platform Enablement

Begin with activating your Salesforce org to host agentic components and ensuring your licenses are recognized across the platform. While Data 360 is typically provisioned automatically as soon as a license is added to your org, you may need to enable it manually if you’re working in a Developer Edition or a sandbox environment. For detailed steps, see Set Up and Turn on Data 360.

Next, you need to enable the Einstein Trust Layer to enforce enterprise-grade privacy and security boundaries.

Permissions

Before you configure agentic analytics, you must define your team’s permission sets. Managing these permissions requires a System Administrator user profile or Salesforce Setup access. While you can customize access, most agentic implementations rely on these permission sets.

Permission Set	Strategic Scope
Data Cloud Architect	The primary builder of the grounding layer (typically, the admin). They manage data mapping, identity resolution, and the creation of the unified profiles used by AI agents.
Data Cloud User	Views agentic performance and cross-functional insights but can’t change the underlying data architecture.

Even as a System Administrator, you must explicitly assign yourself the Data Cloud Architect permission set to access the Data 360 setup menus and specialized grounding tools.

Connect and Prepare Your Data

To bring your data into Data 360, you must first create connections to external systems and establish data streams. After data is flowing, you can standardize, model, and unify the information so AI agents can easily understand it.

Link your business platforms to Data 360 using connectors, and then create data streams to act as the pathways for the information. This step brings multiple data sources into the platform and ensures that AI agents have an up-to-date, secure, and complete view of your company’s data.

Prepare, Harmonize, and Unify Data

With your data streams active, the next step is to translate that incoming information into a common language that your agent can understand. These steps build a clean foundation and ensure the agent references a single source of truth to deliver accurate, non-contradictory answers.

Prepare data: Standardize inconsistent data across systems. For example, combine Gold Status from your sales database and Tier 1 from your support portal into a single, unified status like High Priority.
Harmonize data: Map your cleaned fields to objects found in Data 360’s standard data model. This acts as a universal translator, grouping different labels (like Customer ID and User Account) under a single profile that the agent can easily navigate.
Unify data: Use customizable match rules to link fragmented records belonging to the same person. For example, this process helps the agent see that J. Smith and Jane Smith are the same individual, which provides a complete view of the customer.

To learn more, see Prepare and Model Data and Unify Source Profiles.

Set Up Search Indexing for AI Retrieval

After unifying your data in Data 360, you must generate a search index to make the data discoverable by your agent. This index acts as an organized catalog, scanning your data fields and unstructured documents so the agent can look up millions of records instantly.

To maximize retrieval accuracy, you can use different search index types, parsing methods, and chunking strategies. Use Intelligent Context in Data 360 to process unstructured data and create search index configurations tailored to your specific business contexts.

Move to the Next Layer

Your data is now accurate, indexed, and ready to ground your agent, but it still lacks human context. To give your data a logical framework that understands your business terminology, you must configure the next layer: semantics.

예상 시간

주제

도움말 검색