Microsoft and Salesforce launch Copilotforce to automate CRM actions
The new integration allows Microsoft Copilot to perform tasks directly within Salesforce, moving toward an AI-driven interface for business software.
The short version
- The 'Salesforce in Copilot' plugin includes over 38 pre-configured sales skills.
- Users can link Salesforce data with Outlook emails, Teams meetings, and SharePoint documents.
- The integration uses the Model Context Protocol (MCP) for authenticated access to Salesforce.
- A 'Zero-Copy' method allows data analysis across platforms without creating duplicate copies.
- Gartner estimates the market potential for this shift to AI-driven CRM interfaces could reach $109 billion by 2030.
Why it matters
This marks a shift from humans navigating complex software menus to AI agents performing actions on their behalf. It suggests that the primary way people interact with business tools may soon be through a conversational interface rather than a dashboard.
A new way to work
Microsoft and Salesforce have announced a deepened partnership called Copilotforce. This initiative aims to integrate Salesforce customer relationship management (CRM) functions and business processes directly into Microsoft Copilot. Instead of just searching for information, users will be able to trigger specific actions within the CRM system using the AI assistant.
The core of this announcement is the 'Salesforce in Copilot' plugin, which is currently available in beta. This tool includes more than 38 pre-configured sales functions, which the companies call 'skills.' These skills allow users to connect Salesforce data with common workplace tools like Outlook emails, Teams meetings, and documents stored in SharePoint or OneDrive.
By linking these different sources of information, the companies hope to help sales teams evaluate opportunities more effectively. For example, the system can help identify risks during contract renewals or prepare follow-up activities more quickly. This integration is intended to streamline how employees interact with complex business data on a daily basis.
The technical foundation
The technical backbone of Copilotforce relies on Microsoft's context layer known as Work IQ. This system works alongside the Model Context Protocol (MCP) to provide authenticated access to Salesforce. This means the AI can securely reach into the CRM to pull data or execute tasks while following specific business rules.
One of the key technical highlights is the use of a 'Zero-Copy' process for data exchange. This method allows supported data sets to be queried and analyzed across platforms without the need to copy them into a second system. While the ability for Salesforce Data 360 to access Microsoft's OneLake is already available, a beta version is being released for the reverse direction.
Security remains a central component of this architecture. Even though the AI can perform actions, the access rights and business rules are still managed by Salesforce. Furthermore, any actions that modify CRM records require the user to provide explicit confirmation before the change is finalized.
Expanding the ecosystem
While this partnership is significant, it is not the only move Microsoft is making in the enterprise space. The company already has similar collaborations with SAP, ServiceNow, and Workday. These partnerships allow Copilot to connect with various corporate applications and their respective AI assistants, such as SAP Joule or Workday's Sana platform.
Salesforce is also pursuing its own multi-platform strategy. The company has expanded its partnership with OpenAI to bring Agentforce features to ChatGPT. Additionally, Salesforce has established integrations with Anthropic's Claude and Google Cloud's Gemini Enterprise to connect Salesforce functions to those specific AI models.
Despite these deep collaborations, the two companies remain competitors in the software market. Microsoft continues to offer its own CRM platform, Dynamics 365. At the recent Microsoft Ignite event, the company showcased new Copilot integrations for Dynamics 365 and even introduced tools specifically designed to help customers migrate from Salesforce to Microsoft's own system.
The shift to 'headless' software
Copilotforce represents a broader architectural shift in how business software is designed. Experts describe a move away from traditional user interfaces—which rely on menus, dashboards, and forms—toward AI assistants that interpret requests and execute actions through backend interfaces. This trend is sometimes referred to as 'headless' software, where the underlying system handles the data and rules while the AI acts as the primary interface.
Industry analysts are watching this transition closely as it changes the relationship between software providers and users. If AI agents become the primary way people interact with tools, the actual software application might become 'invisible' to the end user. In this scenario, the user's primary relationship shifts from the software provider to the operator of the AI platform.
The financial scale of this shift is substantial. Gartner has analyzed the market potential for this transition in CRM systems, estimating it could reach up to $109 billion by 2030. This move could turn traditional SaaS applications into what some call 'featureware,' where the software provides the capabilities but the AI provides the experience.
What to watch
- How many enterprise users will actually switch to AI-driven interfaces versus traditional dashboards?
- Whether other CRM competitors will release similar 'headless' integrations to stay relevant.
- The full rollout of the Zero-Copy data exchange between Salesforce and Microsoft Fabric.
Try it
Check if your organization has the 'Salesforce in Copilot' plugin available in the Microsoft Copilot store to see which of the 38 pre-configured skills are enabled for your team.
Sources and links
- OfficialGartner spricht gartner.com
- Officialein MCP-Plug-in für Copilot als Early-Access-Version freigegeben blog.workday.com
- OfficialKern der Ankündigung news.microsoft.com
- Reportingheise.de: the article we sourced
First reported by heise.de. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.