This browser is no longer supported.
Upgrade to Microsoft Edge to take advantage of the latest features, security updates, and technical support.
A main support agent has a connected escalation specialist with a clear routing description. During evaluation, the main agent refuses an escalation request instead of delegating it. What should the maker investigate first?
Whether the expected response exactly matches the wording produced by the connected agent
Whether the main agent's instructions conflict with or block the connected agent's escalation domain
Whether the connected agent has the same display name as the main agent
A support agent uses user authentication to retrieve SharePoint documents. Retrieval works for the maker on the Preview tab but returns errors for another user in Microsoft Teams. What is the most likely cause?
The Teams channel removes SharePoint knowledge source configurations during the publish process
The published user's account doesn't have permission to access the SharePoint site that the agent queries
SharePoint knowledge sources are only available on the Evaluate tab and not in published channels
An organization is preparing to move an agent from a development environment to production. Which practice most directly reduces the risk of an untested configuration change affecting production users?
Increasing the Copilot Credit allocation in the production environment before deploying
Reviewing the agent's past activity on the Monitor tab immediately before deploying
Packaging the agent in a solution, testing it in a separate environment, and deploying the tested solution to production through a controlled process
A maker wants to reduce unnecessary Copilot Credit consumption during the development of a new agent feature. Which approach achieves this goal?
Using the Preview tab for exploratory testing during development and running structured evaluations only at key milestones, such as before initial release and after significant changes
Saving the agent's instructions without publishing to avoid any model invocations during authoring
Deleting and recreating the agent for each new feature to reset the credit counter
You must answer all questions before checking your work.
Was this page helpful?
Need help with this topic?
Want to try using Ask Learn to clarify or guide you through this topic?