Salesforce and Anthropic have announced Claudeforce, an expanded partnership intended to make Salesforce data, business logic, workflows, and governed actions available within Claude.

The first product in the partnership is Salesforce in Claude, a plugin with 37 prebuilt sales skills. Salesforce said the plugin can give sellers and agents access to live revenue context for tasks such as meeting preparation, deal health review, and pipeline review, while allowing approved actions to be taken from within Claude.

The companies describe the integration as a way to bring a reasoning model closer to the customer information and workflow rules that already sit inside a CRM. Salesforce said its AIforce architecture uses Model Context Protocol servers, APIs, and command-line tools to expose approved business data and workflows to agents.

Salesforce and Anthropic also plan to expand integrations across Claude, Salesforce, and Slack. The release said select pilots are available now, with an open beta expected in September.

The announcement arrives as enterprise software vendors race to connect generative AI assistants to operational systems. The practical question for customers will be how permissions, data access, auditability, and human approval are configured when an assistant can move beyond retrieving information and begin initiating workflow actions.

Salesforce launches startup program in Malaysia and the Philippines