Currently available to select pilot customers, the product is scheduled for an open beta in September 2026.
This shift addresses a growing industry move toward "agentic interfaces," where AI agents handle complex workflows that previously required thousands of manual clicks.
By turning its CRM into a "headless" infrastructure—meaning the underlying data and logic are accessible without the traditional visual dashboard—Salesforce aims to increase productivity for knowledge workers.
According to Salesforce executives, the goal is to unlock "trapped value" by allowing AI to synthesize information and execute tasks in seconds, transitioning the business model from a per-seat license toward API consumption, or charging based on the volume of data calls made by the AI.
The partnership relies on a technical architecture that uses Model Context Protocol (MCP) servers to ensure the AI respects existing enterprise security and user permissions.
While the initial rollout focuses on sales, Salesforce plans to launch additional skills for marketing, commerce, and service functions starting in the third quarter.
This integration also deepens the alliance between the two companies, as Claude becomes the default model across other Salesforce-owned properties like Slack, further embedding Anthropic’s generative intelligence into daily corporate workflows.