Salesforce has launched Headless 360, redesigning the platform as a backend execution layer rather than a destination you log into.
Having announced the launch at Salesforce TDX, this update ensures that users will no longer need to open Salesforce to get work done, with workflows surfaced directly in tools such as Slack, Teams, WhatsApp, and voice interfaces.
This is not just about removing the browser, but also about decoupling execution from interface as Salesforce becomes an invisible layer that powers workflows everywhere.
Speaking at the keynote on Wednesday, Joe Inzerillo, President of Enterprise and AI Technology at Salesforce, explained how Headless 360 lets companies build both traditional rule-based software and newer AI-driven systems together, without needing deep AI expertise.
“Software, traditionally, has been deterministic. You put in a set of data, you write some procedures, and the same thing comes out the same side every day.
“55% of enterprise companies say that the biggest problem that they have, is that they do not have enough AI talent.
“Today, we are going to show you how you can use Salesforce Headless 360 to change the way that you’re fundamentally developing code, not just the agentic side, but also the deterministic side.”
CRM Actions Without a Central Interface
This shift in how enterprise software is accessed and used on Salesforce relocates the center of activity for its users, moving them away from the traditional requirement to enter a dedicated interface, navigate records, and manually execute workflows.
Now that dependency on a central UI has been removed, every capability, such as updating a case or triggering an approval chain, is now exposed as an API, tool, or command that agents can invoke directly.
Instead of users going to Salesforce, the platform comes to wherever the work is already happening, enabling agents to surface structured, interactive workflows directly inside communication channels such as Slack, WhatsApp, Microsoft Teams, and voice interfaces.
As a result, a support agent handling a conversation does not need to open a CRM tab to update a case as the workflow appears inline, often as a card or interactive element, allowing actions such as approvals, status updates, or rebooking flows to be completed within the conversation itself.
With the conversation now becoming the primary interface, rather than a dashboard or console, work that previously required context switching, such as moving from a chat to a CRM system, is now embedded directly into the communication thread.
For CX, this ensures that the user experience becomes continuous rather than fragmented, whilst agents orchestrate the underlying business logic, workflows, and data, the user interacts only with the surfaced outcome.
By exposing all core capabilities programmatically, Salesforce enables both humans and AI agents to operate on the same system of record without replicating logic elsewhere, reducing duplication and ensuring that approvals, compliance rules, and workflows remain consistent regardless of where they are executed.
The removal of the UI as a required entry point is less about eliminating interfaces and more about decoupling execution from presentation.
Experience Layer: Build Once, Deliver Everywhere
Salesforce has also introduced the Experience Layer as part of its shift toward a headless, API-first model, separating what an agent does from how it is presented to the user.
In earlier systems, workflows and their interfaces were tightly coupled, with a case update screen or approval form existing only inside Salesforce.
The Experience Layer renders the same underlying workflow as a native interaction across multiple environments without rewriting it.
Developers define workflows, logic, and data interactions once within the platform, with the layer then translating these into interactive components suited to different surfaces.
A single approval process might appear as a Slack card, a mobile UI element, a voice prompt, or a chat interface, depending on the channel, with the functionality remaining the same while the presentation adapts to the context.
The layer also supports richer interaction models than plain text, rendering structured elements such as decision tiles, data views, or multi-step workflows for enterprise tasks that require more than conversational input.
This allows them to be expressed as interactive components that are native to the host environment, rather than forcing users back into a centralized UI.
From a systems perspective, this design reduces fragmentation across channels, as supporting Slack, mobile apps, and voice interfaces previously required separate implementations.
The Experience Layer instead acts as a unifying abstraction, ensuring all channels draw from the same underlying workflows and data models, improving consistency and reducing maintenance overhead.
This shifts enterprise software toward a service-oriented interaction model, whilst the platform provides context, workflows, and governance, the Experience Layer handles distribution across surfaces.
Users can interact through the tools they already use, and agents mediate between those tools and the underlying system, providing a flexible interface while keeping logic centralized and consistent.
