Share via


Use your own model when generating responses

Important

Some of the functionality described in this release plan has not been released. Delivery timelines may change and projected functionality may not be released (see Microsoft policy). Learn more: What's new and planned

Enabled for Public preview General availability
Users, automatically Mar 2026 -

Business value

When you replace the default language model with a custom or fine-tuned model for summarization in Copilot Studio, your organization delivers more accurate, compliant, and domain-specific agent responses. This capability drives higher customer satisfaction, operational efficiency, and competitive differentiation.

Feature details

When you bring your own model (BYOM) to response generation, you add your custom model deployed in AI Foundry. This approach gives you more tailored answers that fit specific domains.

First, go to the agent settings. Use the response models list to add a new model. Enter the AI Foundry deployment connection details for your model.

Next, define the model prompt with instructions and global variables. With full control over the prompt, you give the model rich, contextual input to generate accurate responses.

You can also choose whether to include citations.

When you connect the custom model and set up the prompt, you get full control over agent response generation.

Geographic areas

Visit the Explore Feature Geography report for Microsoft Azure areas where this feature is planned or available.

Language availability

Visit the Explore Feature Language report for information on this feature's availability.