All too often, an organization’s processes, policies, and domain expertise aren’t documented; instead, they’re referred to with the “institutional knowledge” euphemism while companies hope the people holding that knowledge don’t leave the organization.
But in the AI era, that lack of documentation has a more immediate impact: it means AI agents can’t use the intelligence built up over time and therefore can’t be applied to automating many critical processes that companies rely on.
Microsoft is moving to close that gap with business “skills” in the Dataverse data management platform; skills are reusable processes built with natural-language instructions that AI agents discover then follow once they are invoked. Skills can benefit three types of stakeholders:
- Employees documenting how their team operates
- Agent builders who need agents to follow operating processes instead of generic instructions
- Admins or managers who need to govern how business knowledge is shared and deployed
A business skill describes a specific process — the step-by-step instructions, the information required, and the business rules that apply. When they’re connected to a Dataverse Model Context Protocol (MCP) server, agents discover relevant skills automatically and use them to complete tasks in keeping with organizational standards. The agents require no custom code, no workflow builder, and no switching between apps.
Previously, if an agent needed to follow a multi-step business process — like resolving a customer service issue on an end-to-end basis — that wasn’t documented, the agent had no context for the ways a team actually handled the work in question. Its response, therefore, would be generic, lacking organizational context and the details needed to complete work.
The same process is followed regardless of how many agents access and use a skill. If a skill is updated, the change applies everywhere automatically, without the need to update any agents. Skills are governed by built-in Dataverse sharing and visibility controls.
Business skills live in Dataverse and they’re managed from the make.powerapps.com development portal. An organization can document its process in natural language or upload existing documentation, share it with the right users, define who can see and edit the skill, and deploy it — all within Power Apps.
Skills are defined once and stored centrally in Dataverse, so any agent — whether it’s running in Copilot Studio, GitHub Copilot, Azure AI Foundry, or any MCP-compatible client – can use them.
If you prefer to develop skills conversationally, the Dataverse MCP Server lets you create and update skills by asking an agent.
Business skills are in public preview in Dataverse.
More AI Agent and Data Insights:
- Microsoft Deepens MCP Support With New Power Apps and Dataverse Servers
- Microsoft Framework Supports Diverse Models, Agent to Drive Complex Workflows
- Why Microsoft’s Support for Third-Party AI Tools Is So Important
- MCP Enablement Brings AI Automation to Dynamics 365 at Vast Scale
Tom Smith
Analyst
Cloud Wars, Agent and Copilot
Related Posts
