Agent Interoperability v.41 release - Review what's new and changed and limitations in Agent Interoperability for the v.41 release.

Automation 360

Review what's new and changed and limitations in Agent Interoperability for the v.41 release.

What's new


AI Governance for MCP Inbound tool calls Automation Anywhere now logs every tool call that a third-party AI assistant makes to the Automation Anywhere MCP server, providing Automation Admins with centralized visibility into each interaction for governance, compliance, and troubleshooting of agentic workflows. All MCP Inbound log entries appear in the new Agents tab under AI > Governance in the Control Room. Entries originating from the same external AI assistant session are automatically chained together, giving a complete end-to-end view of the interaction. Each entry records: The tool called and the time of the call Input variable names, data types, and values passed by the AI assistant The output (response) returned by the MCP server Execution status, run-as user, session ID, and device name All MCP Inbound entries are prefixed with MCP Inbound in the activity name to distinguish them from native AI Agent activity. Governance log data can be exported to CSV for audit and compliance reporting. Note: The Running time column currently displays `--` and does not show execution duration. The name of the device on which the automation ran is not displayed in the AI Governance logs. When viewing process execution details for MCP Inbound tool calls, output variables information is not displayed. Goal based Agent (GBA) tool calls appear in the Agents tab of AI Governance logs; however, their Source is displayed as Agent instead of External AI Agent. View AI Governance logs for MCP Inbound tool calls
Audit log entries for MCP inbound agent connections The Control Room now logs audit events for all agent connection lifecycle operations and bot runs triggered through the MCP inbound channel. Automation Leads and GRC Leads can navigate to Administration > Audit log to review the following new event types: Agent connection created Agent connection updated Agent connection deleted Additionally, when an external AI agent triggers a bot run through an agent connection, the standard bot execution events — Device acquired by user, Bot sent to device, Run bot deployed, and Run bot finished are captured and attributed to that connection. Note: Bot trigger information is not captured in the MCP Inbound agent connection audit logs. Audit log entries for MCP inbound agent connections

What's changed


Select specific user now available to all users when configuring inbound tools: When adding or editing an inbound tool in AI > Agent connections, the Select specific user option in the Run-as users step is now available to all users regardless of bot runner license type. Previously, this option was disabled unless the user held an Unattended Bot Runner license. You can now designate any specific user as the Task Executor for unattended automation execution. If the selected run-as user does not have a default bot-running device, a device pool must be configured on the Device pools step to continue.

Limitations


When MCP Inbound tool executions are in parallel, the AI Governance log can display overlapping variable values from concurrent tool executions. Input and output variables from parallel tool runs are not correctly attributed to their respective tool calls in the governance log.
Interactive human-in-the-loop workflows, where an agent needs to ask questions back to the user through the client (for example, for mid-process inputs or forms), are not yet supported.
AI Governance for MCP Inbound tool calls Output variables not captured: For process executions, the AI Governance logs do not display output variable information. GBA tool call source label: Tool calls from a Generative Bot Agent (GBA) appear in the same logs; however, the Source column displays Agent instead of External AI Agent.


Limitations from previous releases
You must manually refresh or edit the agent connections whenever you update any automation. This includes changes such as enhanced descriptions or when you add, update, or remove any input/output variables.
The DiscoverAutomation tool might not work or behave unexpectedly if a user has access to more than 1,000 folders in the repository. This happens because the API times out when handling so many folders. Besides the DiscoverAutomation tool not working, there are no other problems.
When a user uses the DiscoverAutomation tool on public automations, the system only checks the first 200 automations that the user can access. This is based on the permissions set for them in the automation repository. Any automations beyond the first 200 will not appear in the DiscoverAutomation's tool search results. However, the user can set up these specific automations as individual agent connections and use them as separate automation tools.
Checked-in automations might take a while to be discovered. It can take 5–10 minutes to summarize the automation. This depends on how many tasks are checked in and how long it takes to create the summary for each automation.
You must delete the agent connections manually whenever the associated automation is deleted or when the user's access is revoked. The MCP client might take some time to remove or refresh the deleted or revoked automations from the Tools list because of caching.