Salesforce announced on the 19th a major expansion of its AI agent platform “Headless 360.” The company is introducing the new “Headless 360 MCP Server” and “Data 360 MCP Server,” along with Slackbot integration, developer tools, and more than 100 reusable skills. The rollout extends beyond marketing, sales, service, and commerce to major clouds including MuleSoft, Informatica, and Tableau.

At the core of this expansion is the Headless 360 MCP Server, now available in open beta. Through the Model Context Protocol (MCP), AI agents running on Agentforce, Claude, ChatGPT, Cursor, and other platforms can dynamically discover, understand, and invoke Salesforce capabilities in real time. By recognizing metadata, the server enables agents to understand data relationships, permissions, workflows, validation rules, and governance, inheriting pre-configured context within Salesforce to execute business operations.

The Data 360 MCP Server is now generally available. It exposes approximately 200 Data 360 APIs through open standards. AI agents with appropriate permissions can execute tasks via natural-language instructions, including building semantic models, transforming data, generating computed insights, mapping fields, creating audience segments, and activating campaigns. Pre-built Agent Skills for Data 360 are scheduled for general availability in late August 2026.

The Slackbot MCP client is already generally available, connecting to Salesforce’s platform and various MCP servers as well as more than 20 partner applications. Employees can update opportunity information, retrieve contracts, and initiate workflows directly from the conversation interface.

Salesforce has also made generally available more than 100 reusable Agent Skills that bundle orchestration, validation, resources, and business logic. The Salesforce Multi-framework is now generally available as well, while the “Headless Experience Layer” is offered in open beta.

Availability Status of Key Features

FeatureAvailabilityHeadless 360 MCP ServerOpen betaData 360 MCP ServerGenerally availableSlackbot MCP clientGenerally availableReusable Agent Skills (100+)Generally availableSalesforce Multi-frameworkGenerally availableHeadless Experience LayerOpen betaPre-built Agent Skills for Data 360GA planned for late August 2026

Note: Availability status is as of the announcement date.

The Spread of MCP and Salesforce’s Strategy

MCP is an open protocol introduced by Anthropic, published as a standard specification for connecting AI assistants with data sources and tools. Microsoft has also documented how agents can connect to MCP endpoints, and adoption is accelerating across the industry.

Through Headless 360, Salesforce aims to shift from a traditional application-centric architecture to a model where AI agents and external applications can reuse Salesforce capabilities. By enabling not only its own Agentforce but also competing external AI agents such as Claude and ChatGPT to invoke Salesforce functions, the company appears to be positioning itself to enhance its value as a platform.

The integration of Data 360 and Tableau MCP servers into Slack is particularly notable as a move that brings AI, data, and analytics directly into everyday collaboration tools. Employees can now instruct data analysis and insight generation from within Slack conversations, with expectations for improved operational efficiency.

In the enterprise software market, interest is growing in “agentic automation,” where AI agents autonomously invoke functions across business systems to execute tasks. Salesforce’s expansion represents a significant strategic move in the enterprise AI platform race, opening up a broad range of capabilities—including not only its own clouds but also acquired and integrated product lines such as MuleSoft, Informatica, and Tableau—to external AI systems based on protocol standards.