Apps, portals, copilots, Slack, embedded experiences
The delivery surface changes fast. It should not drag the platform boundary with it.
Page 02
The architecture is about governed reach, not just API exposure
The cleanest way to think about Salesforce Headless 360 is as a service boundary between delivery surfaces and the core Salesforce systems that still own data, action, and policy.
Once this layout clicks, the use-case page becomes less speculative and more practical.
Next: Browse use casesRequest / tool call / event
APIs, tool contracts, policy checks, orchestration hooks
Capabilities are exposed in a way that both products and agents can call cleanly.
Authorized action path
Data, workflows, customer context, service and revenue systems
The systems of record stay central even as the interaction model becomes headless.
Channels stay replaceable
Teams should be able to add a new surface without redesigning the platform contract underneath it.
Policy travels with execution
Identity, authorization, and allowed action scope remain attached to the call path instead of being rebuilt in every client.
Orchestration becomes a first-class concern
Headless delivery works best when workflows, events, and agent actions can chain cleanly across systems.
That architecture matters because the newest Salesforce-facing workloads are not browser-first. Some run inside conversational interfaces. Some run inside internal tools. Some are buried in partner flows or service operations. The point of Headless 360 is that these experiences do not need to fake a user session to stay productive.
In practice, the diagram above creates a cleaner split: surfaces focus on experience, Headless 360 focuses on callable capability and policy, and Salesforce core remains the authoritative system for state and action.