Salesforce and Anthropic are pushing their partnership deeper into everyday enterprise work with Claudeforce, a joint effort designed to let businesses use Salesforce data, workflows and governed actions directly through Claude.
The first product, Salesforce in Claude, is a plugin aimed at sales teams. It launches with 37 prebuilt skills covering tasks such as meeting preparation, deal-health reviews and pipeline analysis. Sellers can also use Claude to update pipeline information and take other actions without switching into the traditional Salesforce interface.
Salesforce said the integration is powered by AIforce, its framework for connecting enterprise data and workflows to agents through MCP servers, APIs and command-line tools. The plugin is available to select pilot customers, with an open beta expected in September 2026. More prebuilt skills are planned later this year.
Claude becomes central to Salesforce products
The partnership also runs in the opposite direction. Claude is now the default reasoning model for Salesforce’s Atlas Reasoning Engine and powers Agentforce Vibes and Agentforce Coworker by default. Claude is also available in Agent Builder.
Salesforce is extending the relationship to Slack, where Claude will serve as the default model for Slackbot and power Claude Tag and Slack Code. For customers in regulated industries, Claude can also be accessed through Amazon Bedrock within Salesforce’s Trust Boundary, according to the companies.
“Probabilistic intelligence alone doesn’t run a company, and deterministic systems don’t reason,” Salesforce Chair and CEO Marc Benioff said.
Anthropic CEO and co-founder Dario Amodei said the integration allows businesses to use Claude with customer information and business context stored in Salesforce.
Why it matters: The bigger shift is happening behind the screen
Claudeforce is more significant than another AI integration because it tests a potentially uncomfortable idea for Salesforce: its application interface may no longer need to be the main way customers use its software.
Salesforce argues that its real advantage is the underlying data, metadata, workflows and governance accumulated over decades. If Claude becomes the interface, Salesforce can still remain the system behind the work.
That strategy could also benefit Anthropic. Claude gets a direct route into sales workflows and valuable enterprise context, while Salesforce gets a powerful reasoning layer that can operate on its platform. If users increasingly work through Claude, Anthropic gains more influence over how employees access enterprise software and where AI usage is consumed.
Salesforce will also need to demonstrate that agent-driven usage generates sufficient value to offset any pressure on traditional software-seat models.
Security and economics remain key questions
That arrangement makes permissions critical. Anthropic said it put significant effort into managing access and that the companies are developing “Enterprise Frontier Safeguards” to protect Salesforce data and keep model behavior within defined boundaries.
There is also a new cost equation. Customers need Salesforce for the underlying data and services while Anthropic supplies the model and its token-based inference. Salesforce expects token consumption from Claudeforce to be meaningful, although still well below development workloads.
The commercial relationship is already substantial. Quartz reported that Salesforce expects to spend about $300 million on Anthropic tokens in 2026 and holds an equity stake in Anthropic.
For Salesforce, Claudeforce is therefore a bet that owning trusted enterprise data and business rules can remain valuable even if employees increasingly interact with that information through another company’s AI interface.
Other News: Nutanix Enterprise AI 2.8 adds new controls for AI agents, allowing organizations to govern which tools and data agents can access while tracking and limiting token spending by team, user, or individual agent.