Answered

Stop default response from the agent

(2) Share

Report

Posted on by badgerbadger

Hi Copilot agent experts, question -
How do I prevent the agent from trying to be extra helpful? I want the response to stop after Risks section, and ask the export question using a question in topic.
Even after explicitly telling the agent to STOP after the core response, it still tags on this line at the end. I want to immediately trigger that topic automatically.
Thanks!

Categories:

Topic creation & management

I have the same question (0)

All responses (4)

Answers (0)

Sort by

Suggested answer

11manish 3,333 on at

Like
a
(1)

Report
Copy link

Link copied!

To prevent the agent from continuing after the “Risks” section, you should not rely on prompt instructions alone.

Instead, strictly constrain the model to output only up to the Risks section and move the follow-up question into a separate topic.

Then explicitly route to that topic after the response. This ensures the response ends cleanly and the next step is controlled by the conversation

flow rather than the model.

Was this reply helpful? Yes No
BhaskarDhone 1,398 Super User 2026 Season 1 on at

Like
a
(1)

Report
Copy link

Link copied!

Try adding below instruction to your agent.

Response Guidelines:

- Do not ask clarifying or follow-up questions unless explicitly requested.

- Do not suggest additional topics or next steps.

- Do not encourage continued conversation.

- Provide factual, concise answers and stop.

Was this reply helpful? Yes No
Suggested answer

Haque 3,653 on at

Like
a
(1)

Report
Copy link

Link copied!
Hi @badgerbadger,

Let's make sure we have covered the following areas to stop our intelligent friend :):

Explicit Prompt or System Instruction: In your agent’s prompt or system instructions, clearly specify that the response must be confined to the content within the "Risks" section only.

For example: “Only provide answers based on the information contained within the ‘Risks’ section. Do not include or infer information from other sections.”

Segmented Knowledge Sources or Topics: Structure your knowledge or topics so that the "Risks" section is isolated as a distinct knowledge source or topic. This way, the agent queries only that segment when responding.

Use Content Filtering or Post-Processing: After the agent generates a response, apply logic (in the agent flow or Power Automate) to trim or filter the answer to only include text from the "Risks" section.

Limit Context Window: When sending context to the model, include only the "Risks" section content, excluding other sections to prevent the model from referencing unrelated parts.

Custom Extraction or Parsing: Pre-extract the "Risks" section content into a separate document or data entity and use that as the sole knowledge source for queries that should be limited to that section.

I am sure some clues I tried to give. If these clues help to resolve the issue brought you by here, please don't forget to check the box Does this answer your question? At the same time, I am pretty sure you have liked the response!

Was this reply helpful? Yes No
badgerbadger 8 on at

Like
a
(0)

Report
Copy link

Link copied!

Hi gents, thanks for your replies.

I managed to somehow remove the follow up question at the end using only instructions, but now it leaves a note instead!

Ideally I wouldn't want that either, but now at least it doesn't look "broken".

Thank you for all your responses.

Was this reply helpful? Yes No