Configuring Approved Models for Agent Responses in Copilot Studio

AB-620 By islever Editorial Last updated

Configure custom prompts to use the Foundry model catalog

Short answer

AC Select an approved model from the Microsoft Foundry model catalog and ensure the reusable instruction template is applied to the agent.

Question

A company uses an agent in Copilot Studio that produces structured responses using a reusable instruction template. The responses follow the required tone and format. During a compliance review, the team discovers that the responses were generated using a model that is not on the approved list.
The company requires that the reusable instruction template continues to be used and the model used for prompt-driven responses meets governance requirements.
The solution must ensure that prompt-driven responses are generated using an approved model.
You need to configure the solution that meets the requirements.
Which two actions should you perform? Each correct answer presents part of the solution.
NOTE: Each correct selection is worth one point.
  1. Select an approved model from the Microsoft Foundry model catalog in the configuration that is used for prompt-driven response generation. Correct Answer
  2. Configure Azure AI Search as the grounding source so the approved model requirement is enforced during response generation.
  3. Ensure the reusable instruction template is the one applied to the agent that generates the structured response. Correct Answer
  4. Publish the agent so that the model used by prompt-driven responses is updated automatically.
  5. Enable citations so generated responses can be traced and comply with governance requirements.
  6. Add a line to the instruction template that requires the agent to use only approved models when generating responses.

Key Takeaways

This question tests the separation of model selection and prompt engineering; the common trap is attempting to restrict model usage via prompt instructions rather than administrative configuration.

Learn how to enforce governance requirements by selecting an approved model from the Microsoft Foundry catalog while maintaining existing reusable instruction templates in Copilot Studio agents.

Common mistake: Candidates often choose option F, believing that adding a constraint to the instruction template will force the agent to use a specific model. This is incorrect because prompts cannot override the underlying model configuration set in the agent settings.

Expert Analysis

Why the Answer Is Correct

The correct actions are A and C. To meet governance requirements, you must explicitly select an approved model from the Microsoft Foundry model catalog (Option A). This ensures the underlying LLM used for prompt-driven responses is compliant. Since the requirement states the reusable instruction template must continue to be used, ensuring it is applied to the agent (Option C) satisfies the business continuity aspect without violating governance.

Why the Other Options Are Wrong

Option B is incorrect because Azure AI Search is a grounding source for data retrieval, not for enforcing model selection or governance policies on the base model. Option D is irrelevant; publishing the agent does not automatically update or validate the model against an approved list. Option E addresses transparency but does not solve the core issue of using an unapproved model. Option F is the most common distractor; prompts cannot control which backend model is invoked—model selection is a configuration setting, not a text instruction.

Community Comment Notes

Community consensus strongly supports rejecting Option F. As user Rami7 noted, 'Prompts cannot control which underlying model is selected,' highlighting that model governance is an administrative task. Another commenter confirmed that 'F - Prompts can't enforce which model is actually used.' The community correctly identifies that model selection is a configuration setting distinct from prompt logic.

Exam Strategy

Always distinguish between what a prompt can do (guide tone, format, logic) and what agent settings control (model choice, tools, grounding). Governance questions typically require administrative configuration changes rather than prompt engineering tweaks.

Frequently Asked Questions

Can I restrict models using the instruction template?

No. Prompts cannot control the underlying model selection. You must configure the approved model in the agent's settings via the Foundry catalog.

Does Azure AI Search enforce model governance?

No. Azure AI Search is used for grounding responses with enterprise data, not for controlling which LLM model processes the prompt.

More AB-620 FAQ →

Discussion 4 contributions

Compiled from public community threads and reader submissions, lightly edited for clarity. Handles and vote tallies are shown as recorded, and comments reflect the views of individual contributors.

From public community threads 4

  1. Rami7 👍 2
    F. Add a line to the instruction template requiring approved models Prompts cannot control which underlying model is selected. Model governance is configured administratively, not through prompt text.
  2. 8bb6cf1 👍 1
    C + F. F - Prompts can't enforce which model is actually used. Model selection is a configuration setting, not a prompt instruction.
  3. nnitivv 👍 1
    You first choose what one of the approved models and then ensure that the reusable instruction is being used.
  4. Rami7 👍 1
    so F is incorrect

Reader comments & corrections

No reader corrections yet — spotted an error or have a note? Share it below.

Log in to comment, report an error, or add a note about this question.

Copilot Studio Generative Answer Disclaimer Strategy

S-grade

Evaluate Copilot Agent Configuration Compliance Status

S-grade

Export Copilot Studio Agent for Production ALM

S-grade

Configuring Copilot Agent Content Control and Formatting

S-grade

← Back to AB-620 Study Guide