Salesforce expands AI agent capabilities via Headless 360 and MCP

Salesforce is expanding its AI capabilities through the introduction of Headless 360 and the integration of Agentforce with the Model Context Protocol (MCP). This architecture allows AI agents—including those from third-party providers like Anthropic, Claude, and ChatGPT—to interact directly with Salesforce data and business logic without requiring users to switch interfaces.

Key features of this rollout include the ability for Slackbot to act as a central command hub, enabling users to execute complex tasks like updating sales opportunities or retrieving real-time Tableau analytics through simple natural language queries. The platform is also introducing over 100 reusable AI skills designed to automate repetitive data modeling, mapping, and transformation tasks.

While Salesforce reported that combined Agentforce and Data Cloud annual recurring revenue approached $1.4 billion in Q3 FY2026, industry analysts note that successful adoption depends on data maturity. For AI agents to provide operational value rather than just efficiency, enterprises must ensure their underlying CRM data is complete, accurate, and integrated with external systems to avoid incorrect automated actions.