Forum Discussion
Fabric Data Agent performance is not the same via Copilot Studio agent
Hi kevinvu99
In my opinion, Structured output usually works better than numbered or bulleted lists when you’re calling FDA through CSA. Lists tend to be seen as guidance, so CSA may rewrite them or skip a step, which explains why your Step 3 works every time in FDA but is hit-or-miss in CSA. CSA often treats 'always respond with' as optional formatting and may drop it while shaping the reply. a way to fix this is to make Step 3 part of the required output itself (for ex, step3: this is a AI response) and, if possible, add a simple example. Thinking of the response as a strict format rather than a checklist usually makes things much more consistent.
Thanks.
Hi kevinvu99
Thank you for reaching out to the Microsoft Fabric Forum Community.
A Fabric Data Agent can behave one way when you test it directly in Fabric, and then act differently when it’s called through Copilot Studio. The main reason is that Copilot Studio doesn’t pass the user’s message through verbatim, it often rewrites, shortens, or interprets the prompt before sending it to the Data Agent. When that happens, some of your critical instructions can get lost, which is why the agent might skip step 3 or produce inconsistent results.
To improve this, it helps to make your FDA instructions very strict and self-contained. Use a fixed, structured output format (for ex: Step 1 Step 2 Step 3) and include clear example prompts and example responses inside the FDA’s instruction block. That way, even if Copilot Studio modifies the user’s query, the Data Agent still understands and follows the full 3-step workflow.
If I misunderstand your needs or you still have problems on it, please feel free to let us know.
Thanks.