Skip to main content Link Menu Expand (external link) Document Search Copy Copied
L300 Extending M365

Task 06: Test connected agent orchestration

Description

You’ll provision the Zava Care orchestrator agent and verify that it correctly delegates tasks to its two worker agents - retrieving live claims data from Zava Claims and current pricing from Zava Procurement - to produce a unified response to a complex, multi-step query.

Success criteria

  • You provisioned the Zava Care agent package to your Microsoft 365 tenant without errors.
  • You opened the Zava Care agent in Copilot Chat and sent the emergency response coordination prompt.
  • You confirmed that the agent’s response drew on both worker agents - live claims data and embedded pricing information - in a single reply.

Key steps

01: Provision the connected agent

  1. In the leftmost pane, select the Microsoft 365 Agents Toolkit (m365atk-icon) icon.

  2. Under the LIFECYCLE section, select Provision.

  3. Select the dev environment.

  4. Wait for provisioning to complete.

02: Test multi-agent workflows

  1. Open Microsoft Edge, then go back to m365.cloud.microsoft/chat

  2. In the leftmost pane, under Agents, select Zava Care.

  3. Test the following orchestrated workflow:

    Complex Workflow : Multi Agent Coordination

     Find urgent claims needing immediate attention and match with emergency contractor pricing

    Notice it uses both agents to generate a response.

  4. Select Confirm through the prompts.

    ei4yxiyn.jpg

Well done! You’ve successfully built Zava Insurance’s Connected Agent orchestration system!

Conclusion

Congratulations! You’ve completed the lab. This achievement represents the culmination of a sophisticated multi-agent architecture that represents the future of enterprise AI systems - specialized, coordinated, and infinitely extensible! 🚀

In this lab you built a complete multi-agent system for Zava Insurance’s claims operations. You started by connecting a declarative agent to a live MCP server and verifying its tools with the MCP Inspector. You then configured a fully branded claims assistant with natural language instructions, before creating a second agent that answers pricing questions from embedded PDF knowledge files. Finally, you built an orchestrator agent that coordinates both specialists into a single, unified conversational experience inside Microsoft 365 Copilot Chat.

Next steps

Take these concepts further with in-depth, hands-on labs in Copilot Developer Camp:

Resources