Details
- Salesforce announced support for Anthropic’s Model Context Protocol (MCP) Apps, launching bi-directional extensions in Claude starting with Slack and expanding to Agentforce 360 on February 2, 2026.
- Involves Salesforce, Anthropic, products like Claude, Slack, and Agentforce 360; quote from Nick Johnston, SVP of Strategic Tech Partnerships at Salesforce.
- MCP Apps enable Claude to securely access Salesforce context, data, and actions, allowing users to search Slack conversations, generate drafts, invoke AI in chats, and soon trigger Agentforce actions while maintaining permissions and security.
- Builds on prior Slack-Claude integration from October 2025 and Anthropic partnership expansions for regulated industries like finance and healthcare, where Claude operates within Salesforce's trust boundary via Amazon Bedrock.
- Enhances enterprise AI interoperability; competitors like OpenAI and Google integrate via custom APIs, but Salesforce-Anthropic emphasizes full containment in virtual private clouds for data-sensitive sectors, with users like RBC Wealth Management and CrowdStrike.
Impact
Salesforce's MCP integration positions Claude as a secure entry point for agentic AI in enterprise workflows, reducing tool-switching and boosting productivity in regulated industries. This deepens Salesforce's edge over rivals like Microsoft and Google by prioritizing governance and trust boundaries. Expect broader adoption as it enables seamless AI actions across Slack and Agentforce, accelerating AI trajectory in CRM and collaboration tools.
