Salesforce and Anthropic are tightening their enterprise AI relationship with Claudeforce, a new partnership that puts Claude deeper into Salesforce’s CRM ecosystem while giving Claude users direct access to Salesforce data, workflows and business actions.
The first product emerging from the expanded partnership is Salesforce in Claude, a plugin built around 37 preconfigured sales skills. Instead of switching between an AI assistant and a CRM dashboard, sales teams will be able to ask Claude to analyse live account and pipeline information, prepare for meetings, assess deal health and update Salesforce records from within the Claude interface.
That sounds like a relatively straightforward integration, but the more important part is what happens behind the interface. Salesforce says actions initiated through Claude are routed through its platform so that existing permissions, business rules and governance requirements continue to apply. Administrators connect the service centrally rather than configuring access separately for every employee.
Salesforce is using its AIforce infrastructure to provide that connection through technologies including MCP servers, APIs and command-line tools. The broader strategy reflects an increasingly important shift in enterprise software: AI assistants are becoming another interface to business applications rather than simply sitting alongside them as standalone chatbots.
Claude is also moving further in the opposite direction, into Salesforce itself. Anthropic’s models are available within Agentforce, including as a reasoning option for Salesforce’s Atlas Reasoning Engine, while Claude also powers Agentforce Vibes and Agentforce Coworker by default. Salesforce customers can access Claude through Amazon Bedrock within the Salesforce Trust Boundary, an arrangement aimed particularly at organisations with tighter security and compliance requirements.
Slack is another major part of the deal. Salesforce says Claude will be the default model across several Slack AI experiences, including Slackbot, while the companies are connecting conversations in Slack with actions performed in Salesforce. The idea is that employees can discuss a customer, deal or task and then trigger relevant CRM workflows without moving into another application.
The relationship extends beyond customer-facing integrations. Salesforce says it will make Claude Code and Claude Enterprise available to its developers and knowledge workers as preferred AI tools, while Anthropic uses Salesforce as its preferred CRM and Slack as an internal collaboration platform. Salesforce claims its internal use of Claude-powered Slackbot is generating 8.1 million hours of annualised productivity gains, although that figure is a company measurement rather than independently verified evidence of savings.
Claudeforce arrives as enterprise AI competition increasingly shifts from model performance alone toward access to proprietary corporate data and the ability to perform controlled actions. Giving an AI agent permission to modify a sales pipeline is considerably more consequential than letting it summarise one, making governance and reliability central to whether these integrations become genuinely useful.
Salesforce in Claude is currently available to selected pilot customers, with an open beta planned for September 2026. Additional prebuilt skills are expected to begin arriving during the third quarter.
