Designing Advanced Agent Conversations and Responses
In this section, learners build the skills needed to create richer, more useful agent interactions in Copilot Studio. They learn how to format agent responses, use Adaptive Cards, collect structured input, call tools and flows from topics, and generate AI-powered answers using knowledge sources and prompts.
Lessons
Deliver rich agent responses using Adaptive Cards in Microsoft Copilot Studio
- Configure formatted responses using text, images, videos, and quick replies
- Design informational Adaptive Cards for agent responses
- Use interactive Adaptive Cards to collect structured user input
- Customise cards using the built-in designer and JSON
- Apply channel-specific design considerations for Teams, web, and other surfaces
Take action from agent conversations using topics and tools in Microsoft Copilot Studio
- Add tools to topics to perform real-time actions
- Call agent flows from topics to automate business processes
- Use HTTP request nodes to retrieve and send data
- Handle inputs, outputs, and errors in topic-based actions
- Integrate conversations with external REST APIs
Generate AI-powered agent responses using generative answers in Microsoft Copilot Studio
- Configure generative answers in agent topics
- Connect custom knowledge sources
- Shape AI responses with custom instructions
- Author and configure custom prompts
- Connect prompts to models from the Microsoft Foundry model catalog
Building Multi-Agent Solutions in Copilot Studio
In this section, learners design and build multi-agent solutions that support more complex enterprise scenarios. They learn how to evaluate when a multi-agent architecture is appropriate, use child agents, connect existing agents, and integrate cross-platform agents using Agent2Agent protocol.
Lessons
Design multi-agent solutions in Microsoft Copilot Studio
- Compare single-agent and multi-agent architectures
- Evaluate when to use multi-agent solutions
- Design solutions using child agents or connected agents
- Select the right connection option for the scenario
- Choose orchestration and coordination patterns
Delegate agent tasks using child agents in Copilot Studio
- Describe how child agents support complex agent design
- Create and configure child agents
- Define child agent names, descriptions, and instructions
- Configure inputs and outputs between parent and child agents
- Manage child agents by enabling, disabling, redirecting, or deleting them
Build multi-agent solutions using connected agents in Copilot Studio
- Connect to existing Copilot Studio agents
- Connect Microsoft Foundry agents to a Copilot Studio orchestrator
- Connect Microsoft Fabric Data agents for natural language data queries
- Manage connected agents
- Test connected agent solutions for accurate orchestration
Build cross-platform multi-agent solutions using the Agent2Agent protocol
- Explain Agent2Agent protocol and its role in agent integration
- Compare A2A with MCP and other integration options
- Connect Copilot Studio agents to external A2A-enabled agents
- Configure authentication for A2A connections
- Validate agent delegation and review A2A message payloads
Integrating Agents with Enterprise Systems
In this section, learners connect Copilot Studio agents to enterprise systems, tools, services, and knowledge sources. They learn how to evaluate integration patterns, configure connectors and APIs, ground agents with enterprise knowledge, and connect agents to external systems using MCP.
Lessons
Design integration strategies for agents in Microsoft Copilot Studio
- Compare tools, knowledge sources, and agents as integration options
- Evaluate integration approaches for enterprise scenarios
- Identify authentication models across integration types
- Explore the Copilot Studio integration landscape
- Select integration patterns based on business and technical requirements
Take action in external systems using connector and REST API agent tools
- Add prebuilt Power Platform connectors as agent tools
- Add existing custom connectors as tools
- Configure maker-provided and user-provided authentication
- Add REST API tools using OpenAPI specifications
- Configure authentication for REST API tools
Ground agents with enterprise knowledge using connectors and Azure AI Search
- Compare Copilot connectors and Power Platform connectors as knowledge sources
- Add Copilot connectors to agents
- Configure Power Platform connectors as real-time knowledge sources
- Connect agents to Azure AI Search vector indexes
- Configure authentication for knowledge-grounded publishing
Integrate agents with external systems via MCP in Microsoft Copilot Studio
- Explain how Model Context Protocol supports standardised tool integration
- Connect agents to existing MCP servers
- Configure API key and OAuth 2.0 authentication
- Enable and manage MCP tools
- Determine when to use MCP instead of connectors or REST APIs
