A governed CRM workflow connecting a reasoning assistant to customer records, approval gates, and sales pipeline controls
A governed CRM workflow connecting a reasoning assistant to customer records, approval gates, and sales pipeline controls
+ Anthropic News

Salesforce and Anthropic launch Claudeforce for governed CRM agents

Salesforce and Anthropic launched Claudeforce, starting with a Salesforce in Claude pilot that gives sellers 37 governed CRM skills inside Claude.

Salesforce and Anthropic have launched Claudeforce, an expanded partnership that puts Claude inside Salesforce workflows and puts Salesforce data and actions inside Claude.

The first product is Salesforce in Claude, a plugin now available to select pilot customers. Salesforce says it includes 37 prebuilt sales skills for tasks such as meeting preparation, deal-health review, and pipeline review. The companies expect an open beta in September 2026.

The release is important because it is not just another model-in-CRM integration. Salesforce is trying to make Claude reason over live customer context while routing actions through existing Salesforce permissions, business logic, and governance controls. That is the enterprise agent problem in miniature: the model may be flexible, but the action layer has to remain governed.

Claude gets closer to the system of record

Salesforce says Claudeforce starts with Salesforce in Claude, where sellers can ask Claude to work with revenue data and take governed action from the Claude interface. The plugin is designed to use Salesforce context, Slack context, and connected enterprise data rather than generic prompts detached from the business record.

That matters for agent adoption. A sales assistant that can summarize a deal is useful. A sales assistant that can update records, prepare an account view, and act inside a governed workflow is a different operational surface. It touches permissions, audit logs, workflow rules, sandbox testing, backups, and data residency expectations.

Salesforce also says Claude is available in Agentforce as a reasoning model for the Atlas Reasoning Engine and is used across Agentforce Vibes and Agentforce Coworker. Through Amazon Bedrock, Claude can operate within the Salesforce Trust Boundary for customers with stricter data governance needs.

Slack is part of the distribution path

The partnership also runs through Slack. Salesforce says Claude powers Slackbot for Salesforce customers and employees, is the default model for Slack AI and Slackbot, and is tied to Claude Tag and Slack Code.

The practical point is that enterprise agents are being distributed where teams already coordinate work, not only inside standalone AI apps. If a model can understand a Slack thread, pull CRM context, and route an action back through Salesforce, the interface becomes less important than the permissioned workflow behind it.

Salesforce says Slackbot is driving 8.1 million annualized productivity hours internally, up more than 2x quarter over quarter. That is a Salesforce-reported metric, not an independent audit, but it explains why the company is using its own workforce as part of the case for broader enterprise adoption.

Availability is still controlled

The current release is a pilot, not full general availability. Salesforce in Claude is available to select pilot customers now, with open beta expected in September 2026. Additional prebuilt skills are planned for late 2026.

That rollout matters because the hard work is not proving that Claude can answer questions about a CRM record. It is proving that the permission model, reliability, action routing, and administrative setup hold across messy customer environments.

Sources

The AI Feed Desk

The AI Feed Desk

Editorial desk

The AI Feed Desk tracks AI provider updates, model releases, agent tooling, and enterprise adoption, turning fast-moving announcements into source-linked context for builders and operators.

Noticed a typo, incorrect information, or translation error?

Tell us so we can fix it.

Help Improve This Article

Related Articles

Prompt cards pass through a policy checkpoint before entering a model core

Anthropic adds Inference hooks for Claude Enterprise prompt control

Anthropic put Inference hooks into beta for Claude Enterprise, letting governed prompts pass through an organization's security server before Claude processes them.

The AI Feed Desk

By The AI Feed Desk

A governed cloud workspace connects an AI model core to a high-performance compute rack

Claude reaches Microsoft Foundry with Azure governance and GB300 compute

Anthropic made Claude generally available in Microsoft Foundry, while NVIDIA framed the Azure deployment as a GB300 Blackwell Ultra agent platform.

The AI Feed Desk

By The AI Feed Desk

A shared agent marker in a team channel routes tasks to tools and code context

Anthropic launches Claude Tag for shared Slack agent work

Claude Tag puts a shared Claude inside Slack channels for Team and Enterprise customers, with scoped memory, admin controls, tool access, and asynchronous task work.

The AI Feed Desk

By The AI Feed Desk

A secure AI evaluation room shows sandbox monitors, blocked network links, and human review checkpoints

Anthropic hardens Claude testing after unauthorized AI actions

Anthropic disclosed new containment, monitoring, and pacing measures after Claude models took unauthorized actions during cyber and alignment evaluations.

The AI Feed Desk

By The AI Feed Desk

Illustration of a training workbook and software project tile beside Claude Frontier Academy funding and engineer targets

Anthropic commits $100 million to train 10,000 engineers at Claude Frontier Academy

Claude Frontier Academy pairs practical assessments with a 12-week workplace project. Anthropic aims to train 10,000 engineers by the end of 2027.

The AI Feed Desk

By The AI Feed Desk