Enable Copilot chat for app users

Completed

Copilot chat in model-driven apps gives app users an AI-powered chat pane they can use to ask questions about their data in natural language, without needing to know how the app is structured or which views to open.

What can Copilot chat do?

When enabled, Copilot chat appears as an icon in the right navigation bar of the model-driven app. Users can open the chat pane and type questions about the data in their Dataverse environment — for example, "Show me all active cases created this week" or "How many open opportunities do I have?" Copilot answers with data from the tables the user has permission to see.

Copilot can also help users navigate the app. If a user types "Take me to online cases," Copilot automatically opens the relevant screen.

Note

Copilot chat is read-only. Users can retrieve and explore data through Copilot, but they can't create or update records through it.

Enable Copilot chat

Copilot chat for model-driven apps isn't enabled by default. A Power Platform administrator must turn it on for each environment.

  1. Sign in to the Power Platform admin center.

  2. Go to Manage > Environments and select the environment. Select Settings on the command bar.

  3. Select Product > Features.

  4. Set Allow users to analyze data using an AI-powered chat experience in canvas and model-driven apps to On, then Save.

Disable Copilot chat for a specific app

Even when Copilot chat is enabled at the environment level, a maker can turn it off for a specific app.

  1. In the app designer, select Settings on the command bar.

  2. Select the Upcoming tab, set Copilot control to Off (or Default to follow the environment setting), then Save.

  3. Save and publish the app for the change to take effect.

Important

Copilot chat in Power Apps is being deprecated for model-driven apps in environments that are not enabled for Dynamics 365 apps. For those environments, Microsoft recommends switching to Microsoft 365 Copilot in model-driven apps. Dynamics 365-enabled environments are not affected.

Next, you'll learn how to add autonomous agents and an app assistant agent to your model-driven app.