Anthropic’s Dario Amodei Expands Claude-Salesforce Deal Amid AI Safety Debate

Anthropic CEO Dario Amodei is facing renewed attention over the contrasting roles of advanced AI in his company’s business strategy. While he recently called for a slower and more carefully managed pace of frontier AI development, Anthropic has expanded its partnership with Salesforce to integrate Claude directly into enterprise software and workflows.

The collaboration, called Claudeforce, brings Claude into Salesforce, Agentforce and Slack. A major product from the partnership, Salesforce in Claude, is available in beta with 37 pre-built sales skills.

The integration allows sales teams to bring customer accounts, opportunities and pipeline information into Claude conversations. Users can ask the AI to prepare for customer meetings, review deals, analyse pipelines, create dashboards and generate forecasts without switching between multiple Salesforce records and reports.

Claude is also being integrated into Salesforce’s Agentforce platform, where it can provide reasoning capabilities for the Atlas Reasoning Engine and support Agentforce Vibes and Agentforce Coworker. In Slack, Claude is being used for Slack AI and Slackbot, connecting workplace conversations with governed Salesforce actions.

Salesforce CEO Marc Benioff has described the arrangement as combining Claude’s probabilistic AI capabilities with Salesforce’s structured data, workflows and business rules. Amodei has similarly said that the partnership allows businesses to use Claude with the customer information and operational context stored in Salesforce.

The deal comes amid debate over whether AI agents could eventually disrupt traditional software-as-a-service businesses. Amodei has previously rejected the idea that Anthropic’s objective is to eliminate SaaS companies, saying AI agents and existing software can work together.

At the same time, Amodei has called for greater caution in developing frontier AI. In his recent essay, We Must Pace the Frontier, he argued that AI capabilities are advancing rapidly and raised concerns about systems potentially contributing to the development of future AI models.

His proposal does not call for stopping AI development entirely. Instead, he has advocated a more balanced pace alongside stronger alignment, security and testing measures. His suggested approach includes independent third-party evaluations, cooperation among frontier AI companies in democratic countries and eventually broader international coordination.

The Salesforce partnership therefore illustrates Anthropic’s continued push to deploy Claude across business software while its CEO simultaneously argues for greater caution around the pace and safety of frontier AI development.