How Do Customer 360 and Data 360 Connect Data Across a Business?

Updated

Customer 360 and Data 360 split the data problem in two. Data 360 federates, integrates, and harmonizes data from across the enterprise so it is consistent and current. Customer 360 sits above it as the application layer of business rules, workflows, semantics, and operating logic — the layer that tells AI not just what the data says but how the business actually works. Together they produce a single live customer record that both people and AI agents can act on.

What each layer does

Data 360: the trusted context layer

Salesforce describes Data 360 as federating, integrating, and harmonizing enterprise data so agents have trusted context to work from. The practical meaning of those three verbs:

  • Federate — reach data where it already lives rather than forcing every source into one store first.
  • Integrate — connect sources so records about the same customer resolve together.
  • Harmonize — reconcile formats, definitions, and duplicates so "customer" means the same thing everywhere.

The output is a unified, real-time view: Salesforce's page frames it as "Every source, unified in real time."

Customer 360: the operating-logic layer

Customer 360 is not a second database. Salesforce positions it as the application layer holding business rules, workflows, semantics, and operating logic. That is what makes a unified record usable: an agent can see not only that a customer has an open renewal, but which approval steps, entitlements, and team ownership rules apply to it.

How the two connect

The flow runs in one direction, with feedback:

  1. Source systems (sales, service, marketing, commerce, finance, and external systems) hold raw records.
  2. Data 360 federates and harmonizes those records into a trusted, real-time customer view.
  3. Customer 360 applies business rules and workflow logic to that view, producing the single record teams and agents work from.
  4. Activity generated in Customer 360 — cases closed, deals advanced, campaigns run — flows back as new data, keeping the record live.

Salesforce's own framing of the result is "One record, every customer, see it live."

Why this matters for AI agents

Agentforce agents handle research, validation, and repetitive work inside guardrails the business controls. That only works if the context they draw on is accurate and governed. Salesforce states the dependency directly: Data 360 exists so agents have trusted context, and Customer 360 is how AI understands not just the business data but how the business actually works.

A concrete example: a renewal agent needs the contract date, recent support history, current usage, and the discount-approval rule. The first three come through Data 360; the approval rule comes from Customer 360. Miss either layer and the agent either lacks facts or lacks permission to act correctly.

What to check before relying on this architecture

  • Source coverage. Confirm which of your systems Data 360 can federate versus which need integration work first.
  • Definition ownership. Harmonized data is only as good as the agreed definitions; identify who owns customer, account, and product master data.
  • Rule placement. Decide which business rules belong in Customer 360 workflows versus remaining in source systems.
  • Agent guardrails. Map what agents may read and act on before enabling end-to-end resolution.

Pricing and edition details for the underlying products are published on Salesforce's separate product pricing pages; evaluate those against your source-system landscape rather than assuming a single bundle covers every connector you need.