As AI agents take on more autonomous work inside business applications, the question shifts from "can the agent do it?" to "how does a user stay in control?" This session introduces the Power Apps MCP Server and Agent Feed, a public preview feature in Power Platform that lets autonomous agents built in Copilot Studio surface tasks, flag blockers, and propose data changes directly inside model-driven apps, keeping business users in the loop without breaking the agent's flow. Jacob Wilkinson from Microsoft walks through how it works, how to set it up, and a live demo of an agent and human collaborating in real time.

✅ Chapters:
00:00 Introduction
01:03 AI vision and problem context
04:41 Power Apps Agent Feed Overview
05:22 MCP Server & Key Tools Explained
06:38 Demo 1: Agent Feed + Claim Intake Agent
10:09 Demo 2: Coverage Determination Agent
16:30 Final Agent Feed Results & Outcomes
17:43 Providing feedback and wrap-up


✅ Resources:
Public preview: Power Apps MCP and enhanced agent feed for your business applications - Microsoft Power Platform Blog 
https://www.microsoft.com/en-us/power-platform/blog/power-apps/public-preview-power-apps-mcp-and-enhanced-agent-feed-for-your-business-applications/

Supervise agents in model-driven apps with agent feed - Power Apps | Microsoft Learn 
https://learn.microsoft.com/en-us/power-apps/user/supervise-agents-with-agent-feed

Add agents to your model-driven app - Power Apps | Microsoft Learn 
https://learn.microsoft.com/en-us/power-apps/maker/model-driven-apps/add-agents-to-app

Work with Power Apps MCP server - Power Apps | Microsoft Learn 
https://learn.microsoft.com/en-us/power-apps/maker/model-driven-apps/power-apps-mcp-server

https://www.linkedin.com/in/wilkinson-jacob

Microsoft Developer

A demo walkthrough of the Power Apps MCP server and Agent Feed, Microsoft's approach to human-in-the-loop supervision of autonomous AI agents. The Power Apps MCP server provides three tools: 'log for review' for passive agent activity monitoring, 'request assistance' for agents to escalate to human users, and 'invoke data entry' for creating Dataverse records from unstructured text. The demo uses an insurance claims scenario with two agents — a claim intake agent and a coverage determination agent — showing how business users can monitor, intervene, and collaborate with autonomous agents directly within model-driven Power Apps.

Supervise Autonomous Agents with the Power Apps MCP and Agent Feed