Foundry Agent Workflow Unable to Hide Background Agent Output

Abanes, Cameron A 105 Reputation points
2026-08-04T15:09:24.6+00:00

I am currently working in Microsoft Foundry on a customer service agent workflow. I have an orchestration agent, several tool agents, and a response agent. I only want the output of the response agent, but no matter what I do the output of the tool agents and orchestration agent always appear. Previously I was experiencing this problem and so I toggled the "Automatically include agent response as part of the workflow (external) conversation." setting to false on the agents that I didn't want showing (Orchestration and tool agents).

I know that it has that stipulation "If conversationID property identifies the workflow conversation, setting this to false has no impact." underneath, but I don't understand how I can give my agents context of the conversation without giving them the converstationId for the workflow. I have tried adding a "create conversation" node at the beginning of the workflow but it causes agents to lose context between turns so it is less than ideal.

Foundry Tools
Foundry Tools

Formerly known as Azure AI Services or Azure Cognitive Services is a unified collection of prebuilt AI capabilities within the Microsoft Foundry platform

0 comments

1 answer

Sort by: Most helpful
Most helpful Newest Oldest
  1. Allan Solomon Mejia 7,585 Reputation points
    2026-08-04T19:00:15.9966667+00:00

    Hello @Abanes, Cameron A

    The behavior you're seeing is expected when all of the agents are participating in the same workflow conversation. The "Automatically include agent response as part of the workflow (external) conversation" setting only applies when the agent is operating in its own conversation. If the workflow shares a single conversationId, the orchestration and tool agent messages remain part of that conversation context, so setting the option to false won't suppress their outputs.

    If the goal is for the end user to see only the response agent's output, there are a couple of approaches you can consider:

    • Keep the orchestration and tool agents in separate conversations and pass only the required context or outputs to the response agent.
    • Configure the response agent to generate the final user-facing message while treating intermediate agent outputs as internal reasoning or structured inputs.
    • If maintaining conversation continuity is required, verify whether the current Foundry release provides controls for filtering or masking intermediate agent messages in the client/UI. Some behaviors are still evolving as multi-agent orchestration matures.

    From your description, it doesn't sound like you're doing anything incorrectly. It may simply be a current platform limitation when using a shared workflow conversation. If so, it would be helpful if Microsoft could clarify whether hiding intermediate agent responses in a shared conversationId workflow is supported today or is planned for a future update.

    Please "Accept the Answer" if this information helped you. This will help us and others in the community as well.

    Was this answer helpful?

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.