Salesforce connects Claude models to its CRM platform through partnership with Anthropic

Salesforce announced a broad strategic partnership with Anthropic on 27 and 28 August 2026. The collaboration, named Claudeforce, brings Anthropic's Claude models directly into the Salesforce enterprise environment. Sales representatives and AI agents can use it to reason over live customer data and perform pipeline updates without opening the traditional Salesforce interface.

The partnership coincides with strong quarterly figures. Salesforce posted revenue of $11.34 billion in the second fiscal quarter of 2027, an increase of 11 percent year-on-year. The share price rose by approximately 13 percent following the announcements.

The combination of a major technology partnership and upward financial revisions makes this a notable moment for the enterprise AI market, where CRM platforms are increasingly being used as a launchpad for AI agents.

What Claudeforce means for users

The core of Claudeforce is a plug-in called Salesforce in Claude. The plug-in includes 37 pre-built sales skills, ranging from meeting preparation to deal reviews and pipeline analysis. Users can perform actions on Salesforce data directly from the Claude interface, without switching between systems.

Technically, the whole is built on AIforce, a layer that Salesforce describes as an enterprise harness. Through MCP servers, APIs and CLI tools, AIforce makes company data and workflows available to external AI agents without requiring custom integrations. Salesforce positions this as a step towards a headless CRM architecture, in which the value of the platform lies in the underlying data and business logic rather than in the user interface.

Salesforce in Claude is currently available to a select group of pilot customers. An open beta is planned for September 2026, with additional pre-built skills expected in the third quarter of that year. The mutual customer relationship is notable: Anthropic uses Salesforce as its own preferred CRM, while Claude powers the Slack bot for Salesforce customers and employees.

Quarterly figures provide the broader picture

The financial results for the second fiscal quarter of 2027 show how Salesforce's AI deployment translates into measurable growth. Revenue of $11.34 billion was 11 percent higher than a year earlier. Non-GAAP earnings per share came in at $5.90, compared to an analyst consensus of $3.27.

Net income of $3.53 billion was partly boosted by an investment gain of $2.61 billion, largely stemming from Salesforce's stake in Anthropic. This makes the profit figures for this quarter not directly comparable to earlier periods.

On a subscription basis, Agentforce, the platform for autonomous AI agents, grew to more than $1.5 billion in ARR. Combined with Data 360, ARR stands at nearly $3.9 billion, an increase of 210 percent year-on-year. Salesforce revised its revenue guidance for the full fiscal year 2027 to a range of $46.1 billion to $46.4 billion.

Customer usage and internal savings

Salesforce shares a range of concrete usage figures. In the second quarter, customers generated 3.2 billion so-called Agentic Work Units via Agentforce and Slack, 97 percent more than the previous quarter. Over the past twelve months, the total has exceeded 7 billion.

Internally, Salesforce reports that using AI agents in its own operations has saved approximately $100 million in support costs to date, with customer satisfaction remaining the same or improving. A sales development agent processed more than 43,000 inactive leads, which the company says generated $1.7 million in new pipeline. Employees using Agentforce in Slack are said to save an average of five hours per week.

These figures come from Salesforce itself and have not been independently verified. They nonetheless give an indication of the scale at which the company is deploying AI internally and rolling it out to customers.

Dreamforce 2026 as the next step

The annual Dreamforce conference takes place this year from 15 to 17 September 2026 in San Francisco. Salesforce CEO Marc Benioff has announced that the event will focus heavily on the Agentic Enterprise strategy, in which people and AI agents collaborate across multiple systems.

Sessions have specifically been announced on scaling voice AI and connecting content layers to AI agents for better contextual information. For developers and enterprise architects working with Salesforce data, Dreamforce typically provides the first detailed technical explanation of announced functionality. Whether the open beta of Salesforce in Claude will actually go live at that point will become clear in September.