Salesforce has expanded Headless 360 across its platform, extending its AI agent framework across its core software clouds.
The rollout adds new Model Context Protocol servers, Data 360 functions, Slack integrations, developer tools and more than 100 reusable Skills. The aim is to let authorised AI agents discover and use business functions across Salesforce's software estate without separate custom integrations for each new use case.
At the centre of the update is a new Headless 360 MCP Server built on the open Model Context Protocol. It allows AI agents running on platforms including Agentforce, Claude, ChatGPT and Cursor to discover and invoke Salesforce functions in real time while inheriting existing permissions, workflows, validation rules and governance controls.
Salesforce is presenting the shift as a change in how its software is used. Rather than treating each cloud as a standalone application, it is exposing functions from those systems as reusable services that can be called by AI agents, applications and other interfaces.
Data layer
Salesforce is also introducing a Data 360 MCP Server that exposes nearly 200 Data 360 APIs through the same open standard. Authorised agents will be able to build semantic models, transform data, generate calculated insights, inspect identity graphs, create audience segments and activate campaigns through natural language prompts.
The Data 360 MCP Server also includes prebuilt Skills for common workflows, covering areas such as code extensions, transformations and activation tasks.
The announcement extends the framework across a range of product areas. Marketing, sales, service, commerce, data and integration products are being brought into the Headless 360 approach, including Marketing, MuleSoft, Informatica and Tableau.
In marketing, the generally available Marketing Engagement MCP Server allows users to manage journeys, campaigns, data extensions and content through natural language interfaces. In service, headless functions now extend dispatching, field service scheduling and support workflows into external applications and messaging platforms.
In commerce, the model supports AI-driven storefronts and gives developer agents access to commerce functions for building customer experiences. In sales, Salesforce is extending revenue operations beyond the core CRM interface into applications, storefronts and AI buying journeys.
Slack focus
Slack is a key part of the expansion. Slackbot's MCP Client is now generally available, enabling Slackbot to connect with the Salesforce platform, Salesforce MCP servers and more than 20 partner applications, including Atlassian, Box, Canva, Docusign, Notion and Zoom.
Employees can update Salesforce opportunities, retrieve contracts, launch workflows and coordinate work across systems from within Slack conversations. Those actions remain subject to the same identity and governance policies already set by the business.
One partner highlighted the significance of that approach for collaboration software.
"Zoom's open platform strategy is built on meeting customers where they work," said Brendan Ittelson, Chief Ecosystem Officer at Zoom. "By integrating our MCP server with Slackbot, we're making it effortless for millions of users to access their Zoom meeting intelligence through natural conversation - no context switching required. This is what the agentic future of work looks like."
Customer example
Salesforce pointed to travel technology company Engine as an early user. Engine launched its AI support agent, EVA, in 12 days, and the system now resolves half of all customer chat interactions without a human agent.
Another example came from healthcare technology group MIMIT Health AI, which is using Data 360 Headless to access patient information through MCP-based interfaces.
"Data 360 Headless lets our agents reach a unified patient context from any surface - through MCP, without waiting for a purpose-built interface for each one. What makes those answers clinically trustworthy is the physician-curated ontology sitting between the data and the agent. That's where the intelligence lives, and it's why we can hold onto one question, one answer, every time," said Dr. Paramjit "Romi" Chopra, Founder & CEO, MIMIT Health AI.
Developer tools
Salesforce is also expanding the tools available to developers building agent-led applications. Salesforce Multi-framework is now generally available, giving developers a way to build React applications with Salesforce authentication, security and data access built in.
Its Headless Experience Layer is now in open beta for building web, mobile, embedded and conversational applications on top of Salesforce services. Agent Skills and Plugins are also being expanded with more than 100 reusable Skills that package business logic, validation and orchestration into governed components.
The changes reflect a broader effort by software suppliers to make business systems accessible to AI agents through open standards rather than traditional user interfaces alone. Salesforce's latest move places that strategy at the centre of its platform architecture, with MCP servers, reusable Skills and Slack-based actions forming a larger part of how customers can access functions across Marketing, Sales, Service, Commerce, MuleSoft, Informatica and Tableau.