What if the CRM soon no longer needed to be opened?
For decades, using a CRM meant navigating between customer records, dashboards, menus and forms. With Claudeforce, Salesforce and Anthropic want to change this logic: rather than forcing users to go into Salesforce to find information or carry out an action, Claude could become the interface itself, giving direct access to the CRM’s data, workflows and business rules. Announced at the end of August and placed at the heart of the topics expected at Dreamforce 2026, which opens on 15 September, the partnership combines Claude’s reasoning capabilities with Salesforce’s data, business logic, actions and governance.¹ The first concrete product, Salesforce in Claude, allows pilot users to interact with their sales context directly from Claude, without having to keep going back to the Salesforce interface.
Salesforce in Claude starts with sales teams
The first version of Claudeforce targets sales. Salesforce in Claude takes the form of a plugin with 37 preconfigured sales skills, for example to prepare a meeting, review the health of an opportunity or analyse a pipeline.¹ Claude can reason from real-time sales context and carry out governed actions in Salesforce. The integration can also draw on context from Salesforce, Slack and other connectors available in Claude to generate views tailored to the salesperson. The goal is therefore more ambitious than a chatbot able to answer a few questions about a customer: Salesforce and Anthropic want to make Claude an environment from which part of the sales work can be analysed and then carried out. Salesforce points out, however, that Salesforce in Claude is still offered to pilot customers, with a beta announced for September 2026.
Headless 360 gradually separates software from its interface
This integration rests on a major architectural shift at Salesforce: Headless 360. This architecture exposes the platform’s capabilities, including data, workflows, business rules and permissions, so that agents can use them directly through secure connections relying in particular on MCP (Model Context Protocol).² Business software is then no longer necessarily the same thing as its graphical interface. Salesforce can continue to store customer data, apply rules and run processes in the background, while users interact with these capabilities from Claude, Slack or another interface. Salesforce itself sums up this shift as moving from software as an interface to software capable of powering multiple interfaces.
The CRM could become less visible, but much smarter
This shift directly transforms Software and Application Management. In a traditional environment, professionals learn to use an application: they know its menus, fields, screens and procedures. In an agentic environment, they could instead state an objective, such as preparing a meeting with a customer, identifying at-risk opportunities or updating a pipeline, and then let the agent determine which data to consult and which functions to call. The application does not disappear: it becomes a layer of services, data and rules sitting behind the agent. For IT teams, administering business software could therefore gradually mean governing APIs, permissions, agents and MCP connections as much as configuring the interfaces used by employees.
Claude also enters Salesforce, not just the other way round
Claudeforce works in both directions. Salesforce is not simply opening its data to Claude: Claude is also integrated into its own agentic environment. Salesforce states that Claude is available in Agentforce as a reasoning model, that it powers certain experiences by default such as Agentforce Vibes and Agentforce Coworker, and that it is accessible in Agent Builder.¹ The company also states that Claude is the first LLM provider fully integrated within the Salesforce Trust Boundary, notably via Amazon Bedrock. This architecture allows Salesforce to retain its security and governance mechanisms while providing access to the capabilities of Anthropic’s model, a particularly important point for organisations handling sensitive data or operating in regulated sectors.
Slack becomes the other cornerstone of Claudeforce
The partnership takes on another dimension with Slack. Salesforce wants to make its collaboration platform the space where humans and agents work together, and Claude is becoming a central part of this strategy. According to Salesforce, Claude is now the default model for several Slack experiences: it powers Slackbot, takes part in collaboration through Claude Tag and is one of the founding partners of Slack Code.¹ The idea is to allow a team to discuss a situation in Slack, ask Claude to analyse it using the available context and then, where permissions allow, trigger an action in Salesforce without interrupting the workflow. The CRM could then gradually move from being an application you open to becoming business infrastructure accessible from the space where employees already work.
Salesforce permissions follow the user all the way into Claude
Connecting an AI model to a CRM immediately raises a question: what can it actually see and change? Salesforce says that Salesforce in Claude reuses existing permissions and business rules. Claude should only see the information the logged-in user is authorised to access and can only perform actions matching that user’s rights.² This is crucial: a conversational interface must not become a shortcut for bypassing the information system’s security mechanisms. The more agents are able to operate business software directly, the more companies will have to treat identities, permissions, action logs and execution rules as fundamental components of their agentic architecture.
Dreamforce now needs to show how far Salesforce wants to go
The timing is strategic. Dreamforce 2026 takes place from 15 to 17 September in San Francisco, and Salesforce is already announcing demonstrations of Salesforce in Claude as well as new information on what the two companies are building together.³ It is therefore important to distinguish what has already been announced from what is still to come: Claudeforce and Salesforce in Claude do exist, the 37 sales skills have been announced and some customers are taking part in the pilot, but several extensions of the integration have yet to be detailed. Above all, Dreamforce should reveal how far Salesforce intends to push this strategy, in which Claude, Agentforce, Slack, Headless 360 and CRM data function as the different layers of a single agentic environment.
After the graphical interface, the interface could become an intention
Ultimately, Claudeforce illustrates a much broader shift in enterprise software. For a long time, innovation consisted of improving the graphical interface so that users could find a function more quickly. Agentic AI is gradually reversing this logic: the user states what they want to achieve and the agent itself looks for the data, functions and workflows needed to complete the task. Salesforce is therefore not necessarily planning the disappearance of its traditional interfaces, but is preparing for a world in which they would no longer be the only way to access the CRM’s capabilities. If this approach becomes widespread, the next battle in business software may be fought less over the quality of menus than over the ability to expose data and actions securely to AI agents.
360° IT News — decoding technology to understand the jobs of tomorrow.
Further reading
This strategy is a direct extension of a transformation already visible in collaboration tools. Read our CCCLX article “Claude Tag turns Slack into a true agentic workspace”, which analyses how Anthropic is making Slack an environment in which AI can follow teams’ context, join their conversations and take a direct part in their workflows.
References
1. Salesforce. (2026). Salesforce and Anthropic Announce Claudeforce: The #1 AI Meets the #1 AI CRM.
https://www.salesforce.com/news/press-releases/2026/08/27/salesforce-and-anthropic-announce-claudeforce/
2. Salesforce. (2026). Claudeforce: The #1 AI Meets the #1 CRM.
https://www.salesforce.com/claudeforce/
3. Salesforce. (2026). Dreamforce 2026 Official Media Resources & News.
https://www.salesforce.com/news/dreamforce-26-media-resources/
4. ITPro. (2026). What to expect at Dreamforce 2026.
https://www.itpro.com/business/business-strategy/what-to-expect-at-dreamforce-2026
ServiceNow wants to turn process improvement into a continuous AI-driven loop
With AI Workflow Factory, ServiceNow aims to link the identification of inefficient processes with building and running their improvements. The announcement notably includes Autonomous Engineer, a tool designed to automate more of the implementation work.[1] For...
Mistral Large 4: what its trillion-parameter model could change for businesses
Mistral has opened the public preview of Mistral Large 4, nicknamed “Le Chonk”, and plans to release its weights by the end of the month.[1] This European model targets agentic use cases in particular, as well as cybersecurity, finance and industry. For businesses,...
IBM Bob: developing with AI without your code leaving the company
IIBM has announced a self-hosted version of Bob, its agentic development platform, for internal infrastructure, private or sovereign clouds and environments isolated from the internet.[1] This option addresses a concrete obstacle: some applications contain code, data...
HPE is preparing a network that anticipates incidents and automates certain interventions
HPE plans to deploy a Self-Driving Network for the 2027 Ryder Cup, with Wi-Fi 7 and an operations centre bringing together data and AI tools.[1] The stated aim is to monitor usage, adapt the network and move towards autonomous interventions. For businesses, the...
Can your AI agent really defend your company? Hack The Box puts its skills to the test
Hack The Box is launching AI Range Enterprise Edition to let security teams assess their own AI agents on the roles assigned to them.[1] For businesses, the stakes are concrete: having evidence before delegating a task, then checking that this evidence remains valid...
VAST Data wants AI to work with sensitive data in a protected environment
VAST Data has unveiled DataEnclave, a confidential execution capability designed to bring artificial intelligence models closer to companies’ sensitive data. Presented as a preview, it is due to go on sale in the first quarter of 2027.[1] The aim is to make certain...
Databricks buys Row Zero to bring spreadsheets closer to governed data
Databricks has announced the acquisition of Row Zero and plans to integrate its spreadsheet experience into Genie.[1] The aim is to let business teams explore data in a familiar interface, with greater control over access and usage. For companies, the challenge is to...
Google Cloud takes its X5 instances to 48 TB for large SAP databases
Google Cloud has announced the general availability of the X5 series, which raises the maximum advertised memory capacity on a single node to 48 TB, compared with 32 TB for X4.[1] Aimed in particular at very large SAP HANA databases, the offering is based on bare...
Cloud Storage: Google automates the detection of anomalies
A rise in reads, cross-region transfers or stored volume can inflate the cloud bill and disrupt applications. With Storage Intelligence advisor, Google offers a view of unusual changes in Cloud Storage, along with leads for investigating them. For businesses, the...
AI voice agents: Oracle details a LiveKit architecture on OCI
Oracle has described an architecture that pairs the LiveKit Agents framework with Oracle Cloud Infrastructure to automate real-time voice calls. The design links telephony, AI models and business services, while placing internal components in private subnets. For...
