SaaS was never dead. Claudeforce, the newly expanded partnership between Salesforce and Anthropic, is the clearest signal yet. It isn’t AI or SaaS, it’s both.
Claudeforce answers a question enterprises have been wrestling with: not whether to adopt agentic AI, but which model to trust with the data that runs their business. It’s already live with a small group of early customers, putting real CRM data, governed by your own permissions, to work inside Claude. Organisations want AI that reasons over real pipeline, real customer history and real business context, not a chatbot bolted on top of a CRM they still have to open separately. Claudeforce closes that gap directly.
Salesforce in Claude puts live CRM data where sellers already work
The first product out of Claudeforce is Salesforce in Claude, a plugin shipping with 37 prebuilt sales skills covering meeting prep, deal health reviews and pipeline analysis. Sellers, reps working the pipeline, stay inside Claude for the whole job: working live revenue numbers, letting pipeline update itself, and acting on deals with Salesforce’s own rules enforcing every step. Underneath it sits Headless 360, the part of Salesforce built specifically so an AI model can reach the platform’s data, logic and rules directly, connecting once over MCP instead of needing a custom integration for every use case.
The skills go beyond a static dashboard. The first time a seller opens Salesforce in Claude, Claude pieces together everything relevant across their tools, from Salesforce records to Slack conversations, and turns it into a working view of their accounts and pipeline from day one. From there, at speed a seller can:
- Ask what to prioritise today, and get a plan shaped by whatever changed overnight
- Ask about a specific deal, and get the objections you’re likely to hear, and how to answer them
- Ask Claude to flag upsell and renewal opportunities hiding in accounts they already own
Governance sits underneath the whole thing
This is the detail that matters most for trust. Access is not reinvented for Claude. It inherits whatever a given user is already permitted to see and do inside Salesforce, so nothing needs building or auditing separately for the AI layer. Teams determine how much Claude can do unsupervised, from requiring a check before any outbound email to letting it act independently as trust is established, and updates only ever touch the specific field involved. Data submitted through Sonnet, Opus and Haiku within this connection is not retained.
This is bigger than one plugin
Salesforce in Claude is the visible piece, but Claudeforce runs in both directions. Claude is also being built into Salesforce’s own AI stack, sitting behind Agentforce’s reasoning layer and available to regulated industries through Salesforce’s secure hosting arrangement with Amazon Bedrock. Inside Slack, Claude becomes the standard model behind everyday assistant features, in-channel collaboration, and Slack’s new shared coding spaces.
The relationship cuts both ways commercially too. Anthropic now runs on Salesforce as its preferred CRM and Slack as its preferred internal work platform, and Salesforce is rolling Claude Code out to its own developers and knowledge workers as their preferred AI assistant. Each company has become a genuine customer of the other’s technology, using it to take its own product further.
What this means for Salesforce customers
The practical shift is that choosing an AI tool is no longer a one-vender decision. You can run Agentforce’s own AI and Claude cooperatively in the same org, both working under the same Salesforce permissions. That flexibility only pays off if your underlying data and access controls are in good shape first.
That’s where we come in. The hard part hasn’t moved: getting your data and AI readiness sorted, defining the right governance, and building the change management that makes adoption actually stick. It’s still the part that everyone underestimates. At iMMERSIVE, we help clients get that data foundation right, configure Agentforce and connected AI tools accordingly, and keep optimising once it’s live. A partnership like Claudeforce therefore becomes a real capability, not another integration to manage.
Get in touch
Wherever you are in your AI and Salesforce journey, whether you’re still getting your data foundations right or ready to explore what Claudeforce could do for your team, get in touch with one of our team of experts here and we’ll help you work out what this partnership actually means for your org and your data.
