The expanded partnership puts Anthropic's AI closer to Salesforce's customer data and workflows, reflecting a growing push to make enterprise software accessible through AI rather than only through traditional applications. 

[For more news, click here]

A salesperson preparing for a customer meeting can spend as much time gathering context as actually preparing for the conversation. Account history sits in one place, recent messages in another, and notes from previous meetings may change the picture again. AI assistants can already bring some of that information together, but the employee has generally still had to move between the assistant and the software where the work is ultimately recorded. Salesforce and Anthropic are now testing a setup in which the AI can work with that underlying business information and carry out approved actions without requiring the employee to handle every step manually. 

The companies announced their expanded partnership, Claudeforce, this week, with Salesforce in Claude as the first product. The plugin is aimed initially at sales teams and comes with 37 prebuilt skills covering meeting preparation, deal health and pipeline reviews. It can draw on live Salesforce information and initiate actions through the platform, with an open beta expected in September 2026. What makes the arrangement worth watching is less the number of sales functions included at launch than what happens when the AI decides something needs to be done. Salesforce says those actions are routed through its platform, where customer permissions and business rules remain in effect, while authentication is managed centrally. For businesses that have spent years establishing controls around customer information, that distinction matters when an AI moves from providing an answer to changing something inside an operating system. 

The CRM Is Becoming Something Agents Can Work Through 

The move fits a change already underway at Salesforce, its Headless 360 initiative allows AI agents to access capabilities from the company's applications through interfaces including MCP, rather than requiring them to work through the conventional software interface. That means the underlying data and workflows can be used by an agent while the permissions and business rules attached to them remain in place, particularly as companies begin experimenting with AI that can perform tasks rather than simply generate text. 

“We’re bringing together the world's #1 AI and #1 CRM — the best of both worlds. Today, we're launching Claudeforce, running directly on top of Salesforce via our new AIforce UI harness, Headless 360, Data 360, Tableau, and Slack. Here, the UI is the AI — allowing you to build custom apps dynamically and answer any enterprise question. Probabilistic intelligence alone doesn't run a company, and deterministic systems don't reason. By fusing Claude's extraordinary reasoning with the trusted data, workflows, and governance every enterprise runs on, we're delivering a dynamic interface that thinks, reasons, and acts. This is how every business will run.” said Marc Benioff, Chair & CEO, Salesforce 

For Anthropic, access to that underlying business context is an important part of the deal. Customer histories, sales records and established processes give Claude information that would otherwise have to be collected from separate systems or supplied by an employee. Claude is already available as a reasoning model in Agentforce, while Salesforce says the model also powers Slackbot and is being used across other parts of its business. The companies are also customers of one another, making the relationship broader than a single integration between two products. 

"We believe frontier intelligence should be safe, trusted, and deeply capable, which is why the world's leading businesses turn to Claude for their most important work. Salesforce in Claude brings this same frontier intelligence into the systems where much of the world's commercial activity happens. Through this partnership, companies can point Claude at the customer information and business context that they've been building in Salesforce for decades, and use it to actually run and grow their businesses.” said Dario Amodei, CEO & Co-Founder, Anthropic 

Slack is also part of the expanded relationship, with Salesforce saying Claude will become the default model for Slack and power Slackbot, while the companies plan to connect conversations with Salesforce actions. An employee discussing an account or opportunity could retrieve relevant information or complete an approved task without switching applications. The practical test, however, will come with wider use. An assistant that summarizes a customer interaction can still leave the final decision to an employee, while an agent that changes a record or triggers a workflow can have a direct effect on the business. Keeping those actions within existing permissions and rules gives companies a measure of control, but it does not answer whether employees will actually want to hand more of their daily work to an AI system. With Salesforce in Claude still limited to pilot customers, the September beta should offer the first meaningful indication of whether this kind of interaction feels like a useful shortcut or simply another layer between people and the software they already know how to use. 

Related Articles 

Slack to Become AI Hub as Salesforce Tackles Agentic Fragmentation 

AI Agent Governance Is Falling Behind Enterprise Adoption, Optro Research Finds 

Cequence Security Launches Agent Personas to Automate AI Agent Governance Across the Enterprise