Build Power Platform Agents: Foundry vs Copilot Studio
Short answer
Agent1 should be built using Microsoft Foundry to support Semantic Kernel and multi-API connections. Agent2 should be built using Microsoft Copilot Studio to enable direct Dataverse connectivity and embedding in Power Apps.
Question
A company has a Microsoft Power Platform environment.
You need to build two agents named Agent1 and Agent2. The solution must meet the following requirements:
• Agent1 must be extendable by using the Semantic Kernel and must connect to multiple business apps and APIs.
• Agent2 must connect directly to data stored in Microsoft Dataverse and must be embeddable in a Microsoft Power Apps canvas app.
What should you use to build each agent? To answer, select the appropriate options in the answer area.
NOTE: Each correct selection is worth one point.
Key Takeaways
The core trap is confusing 'Copilot in Power Apps' (a UI feature) with 'Copilot Studio' (the agent builder). You must select the platform that natively supports the specific technical constraints like Semantic Kernel or Dataverse embedding.
This question tests the distinction between Microsoft Foundry for extensible, Semantic Kernel-based agents and Copilot Studio for Dataverse-integrated, embeddable agents. The correct configuration assigns Agent1 to Microsoft Foundry and Agent2 to Microsoft Copilot Studio.
Common mistake: Learners often select 'Copilot in Power Apps' for Agent2 because it sounds similar to 'embeddable in Power Apps'. However, Copilot in Power Apps is a user-facing interface for interacting with existing apps, not the development environment for building custom agents that connect to Dataverse.
Expert Analysis
Why the Answer Is Correct
Agent1 requires extension via Semantic Kernel and connections to multiple APIs. This points directly to Microsoft Foundry, which is designed for developer-centric, extensible AI architectures that leverage frameworks like Semantic Kernel for complex orchestration and tool integration. Agent2 must connect to Dataverse and be embeddable in a canvas app. While both Copilot Studio and Copilot in Power Apps relate to this, Microsoft Copilot Studio is the actual authoring tool used to build agents that integrate deeply with Dataverse data models. These built agents can then be embedded into Power Apps canvas applications.
Why Other Candidate Answers Are Wrong
Copilot in Power Apps is incorrect for Agent2 because it is a conversational interface layer within Power Apps, not a platform for building agents. It does not provide the necessary authoring environment to configure Dataverse connections and agent logic. Selecting it confuses the runtime experience with the development tool. Similarly, using standard Copilot Studio for Agent1 would miss the requirement for Semantic Kernel extensibility, as Foundry is the specific enterprise framework for such advanced, code-heavy agent designs.
Community Comment Notes
The community consensus correctly identifies Microsoft Foundry for Agent1 due to its Semantic Kernel capabilities. For Agent2, there was initial confusion between Copilot Studio and Copilot in Power Apps. As one commenter noted, "Copilot Studio... enables building agents that natively connect to Dataverse and can be embedded into Power Apps," clarifying that the builder (Studio) creates the asset that gets embedded (via the interface). Another comment highlighted that "Semantic Kernel... runs outside Copilot Studio" in the context of Foundry, reinforcing the separation of concerns.
Official Reference
https://learn.microsoft.com/en-us/semantic-kernel/
This link opens the official vendor documentation in a new tab.
Exam Strategy
Always distinguish between the agent authoring platform (Copilot Studio, Foundry) and the user interface (Copilot in Power Apps). When a requirement mentions 'embedding', look for the platform that generates the embeddable agent asset, not the viewer.
Frequently Asked Questions
What is the difference between Copilot Studio and Copilot in Power Apps?
When should I use Microsoft Foundry over Copilot Studio?
Discussion 14 contributions
Compiled from public community threads and reader submissions, lightly edited for clarity. Handles and vote tallies are shown as recorded, and comments reflect the views of individual contributors.
From public community threads 14
-
2db2984 👍 15Should be Agent1 -> MS Foundry Agent 2 -> Micorosoft Copilot Studio (enables building agents that natively connect to Dataverse and can be embedded into Power Apps.)
-
hi1011 👍 11Agent 1 -> MS Foundry Agent 2 -> Copilot in Power Apps https://learn.microsoft.com/en-us/power-apps/maker/canvas-apps/microsoft-365-copilot-canvas-app
-
ziggy1117 👍 5Agent 2 - copilot studio is right answer. Copilot in Power Apps is wrong. Copilot in Power Apps lets you quickly build apps using natural language. Describe your business needs, and the service generates an app and data model for you—no coding required. People interact with your app by asking questions or giving commands in everyday language, making the experience more intuitive and engaging. https://learn.microsoft.com/en-us/power-apps/maker/canvas-apps/ai-overview
-
Joselito79 👍 5Explanation Agent1 → Microsoft Foundry: Provides capabilities for building and extending AI agents, including integration with Semantic Kernel and connections to business applications and APIs. Agent2 → Copilot in Power Apps: Designed to work directly with Dataverse and can be embedded within Power Apps canvas apps.
-
omath 👍 3Agent1 → Microsoft Foundry Agent1 must: - Be extendable using Semantic Kernel - Connect to multiple business apps and APIs - Support complex orchestration and multi‑agent patterns These are exactly the capabilities of Microsoft Foundry, which is designed for enterprise‑grade, extensible, multi‑agent AI systems with deep integration into Semantic Kernel. Agent2 → Copilot in Power Apps Why: Agent2 must: - Connect directly to Dataverse - Be embeddable in a Power Apps canvas app This is precisely what Copilot in Power Apps is built for.
Show 9 more comments
-
wrak 👍 3Agent 1 - via Foundry. Semantic Kernel (or the newer Microsoft Agent Framework) runs outside Copilot Studio, as your own agent runtime.
-
Srikantha 👍 3The correct selections for each requirement are: Agent1: Microsoft Foundry Agent2: Microsoft Copilot Studio
-
Mija 👍 3Agent1 → Microsoft Foundry Agent2 → Copilot in Power Apps
-
deca0d4 👍 2Agent1 use Foundry: Use Foundry when you need a more developer-oriented/custom agent architecture, advanced orchestration, Semantic Kernel, custom tools, custom code, or deeper app integration. Semantic Kernel provides an agent framework, and Microsoft documentation describes using Semantic Kernel with Foundry/agents. Agent2 use Copilot for Power Apps: Copilot Studio creates the custom agent. Power Apps lets you add a custom Copilot Studio copilot to a canvas app.
-
Alicia_lili 👍 2Agent1: Microsoft Copilot Studio Agent2: Copilot in Power Apps
-
TheMCT 👍 1Agent1 - MIcrosoft Foundry Agent2 - Copilot in Power Apps Microsoft Foundry -> supports building advanced AI agents that can be extended with Semantic Kernel and integrated with APIs, business applications, and enterprise services. Copilot in Power Apps - > is designed to work with Dataverse data and can be embedded directly into canvas apps for end-user interaction.
-
554b579 👍 1Agent1 Microsoft Foundry — TRUE. Semantic Kernel is a pro-code orchestration SDK, and the Foundry Agent Service is the agent platform designed to be extended with it. Foundry agents also attach tools for APIs and business systems, matching the second half of the requirement.
-
a5f05ea 👍 1Agent1 Extendable using Semantic Kernel + multiple business apps/APIs: Microsoft Foundry Agent2 Direct Dataverse connection + embedded in Power Apps canvas app: Copilot in Power Apps
-
Luweho 👍 1Foundry and Copilot Studio. Will use "Copilot in Power Apps" to integrate the agent in canvas apps, but the question is where to BUILD it.
Reader comments & corrections
No reader corrections yet — spotted an error or have a note? Share it below.