This procedure describes how to set up and use Automation 360 with the Model Context Protocol (MCP) in Microsoft Copilot Studio. By using Agent Interoperability, you can connect several MCP clients to the Automation Anywhere Control Room. This allows MCP clients to search, think, and manage tasks.

It helps users and administrators set up API access in Automation 360, deploy an MCP server, connect it to Copilot Studio, and run bots directly from Microsoft Copilot or Teams. The goal is to allow smooth and secure automation execution and status monitoring within Copilot’s chat interface.

Each MCP client might support different MCP features, allowing for varying levels of integration with MCP servers. For a comprehensive list of supported MCP clients and latest features, click Feature support matrix.

Review this video showing how to configure and run Automation 360 MCP in Microsoft Copilot: Automation 360, MCP, and Microsoft Copilot configuration.

Note: MCP Server configuration is not available in the Control Room right now. Once you set up the Automation Anywhere MCP server endpoint and user credentials in your MCP client, you can use these MCP tools:
  • Discover automations: This MCP tool looks at what the user wants and finds the right automations from the Automation Anywhere public repository. It only gets automations that the user is allowed to use.
  • Run automations: This MCP tool runs Automation Anywhere bots, processes, and API Tasks from third-party AI Agents.
  • Get automation results: This MCP tool gets the status or result of the automation that was started by the Run automations tool.

Procedure

  1. In Copilot Studio, click Agents from the side panel to open the Agent you will use (or create a new Agent).
  2. Scroll down to make sure you set Orchestration - Use generative AI to determine how best to respond to users and events is Enabled.
  3. From the top of the page, click the Tools tab and then click + Add a tool.
  4. In the middle of the Add tool page, click Model Context Protocol and then click + New tool on the right side.
  5. Click the new Model Context Protocol tool you just added and enter the following:
    • Server name: For example, enter Automation 360 MCP Server.
    • Server description: For example, enter Automation 360 MCP Server enables discovering, running, and retrieving results of automations directly through MCP.
    • Server URL: Enter the public URL of your MCP server. For example, https://mcp.stage.acmecompany.com/mcp.
    • Select API key as the authentication type. Then, indicate whether it goes in the Header or Query. (For this example, select Header and enter the Header name X-AUTH.)
    Note: There are two ways you can set up the tools: authentication or the API key of a service user that all the business users of that Agent will use; or switching it to per user mode. If you switch to per user mode, each user who accesses that Agent needs to set the API key once when they access the Agent.
  6. Click Create. The new Automation 360 MCP Server tool (custom connector) has been created.
  7. To connect the new Automation 360 MCP Server tool, click the down arrow next to the Not connected dropdown, and click +Create new connection. Enter your X-AUTH header information and click Create. The new Automation 360 MCP Server tool should now show as connected indicated with a green check mark.
  8. Click Add and configure to link the tool. The tool has now been added to the agent.
  9. At the top of the page, click the Tools tab, select your MCP server (Automation 360 MCP Server), and scroll down to review the tool's: Details, Inputs, Tools, and Resources.
    The Tools section displays a list of available Automation Anywhere tools from the MCP server:
    • Discover automations
    • Run automations
    • Get automation results
  10. From the Test your Agent panel on the right side, click the three horizontal dots, and select Manage connections from the dropdown.
    1. Your MCP server (Automation 360 MCP Server) should display in the list. Click Connect from the Status column.
    2. Click Submit from the Create or pick a connection page. The status changes from Connect to Manage.
  11. Navigate to the Agents page, select Settings->Generative AI. From the Model section, select your agent's model from the dropdown. For this example, select GPT-5 Auto (preview) and click Save.
  12. Navigate back to the Agents->Tools page for your MCP server (Automation 360 MCP Server). In the Details section, click Additional details to view.
  13. In the Test your agent panel, enter a question. For example, enter: Discover list of available automations.
  14. Confirm that the MCP server receives the call, the Automation 360 job executes and completes, and the Agent returns a status or result. Once validated, you can publish the Agent (so end-users can access it through Teams and Microsoft 365 Copilot). All calls create audit events for compliance and debugging.