Salesforce and Anthropic are expanding their partnership to bring Anthropic’s Claude AI deeper into enterprise workflows through Slack and Salesforce’s Agentforce 360, aiming to give businesses AI assistance grounded in trusted data, governance and user permissions.
Under the new integration, Claude can interact with Salesforce data and workflows using Salesforce’s Model Context Protocol framework, allowing users to draft, review, and edit AI-generated content directly in Slack while keeping humans in control. Rather than acting autonomously, AI outputs are designed to remain reviewable, auditable and permission-aware.
Upcoming integrations with Agentforce 360, Data 360, and Customer 360 will allow Claude to invoke Agentforce to generate AI outputs inside Salesforce’s trust boundary, ensuring results are based on secure, governed enterprise data.
Salesforce said the approach helps enterprises move faster without sacrificing accountability, particularly in regulated industries. The Slack app for Claude is available now for Team and Enterprise users, with broader Agentforce integrations rolling out over time.