Here’s a summary of the steps:

Steps in this Quickstart:

1. Create a Project in Azure AI Studio
2. Deploy an Azure OpenAI Model
3. Chat in the Playground Without Using Your Data

Prerequisites:

• Azure Subscription: You’ll need an active Azure subscription. If you don’t have one, you can create one for free.
• Azure AI Studio Hub: You must have access to an Azure AI Studio hub or permission to create one. Your role must be either Azure AI Developer, Contributor, or Owner on the hub.
  • If you’re a Contributor or Owner, you can create a hub during this tutorial.
  • If you’re an Azure AI Developer, the hub must already exist.
• Quota Considerations: Ensure your subscription is within the quota limit to deploy a new model. If not, you should have an already deployed chat model.

Creating a Project in Azure AI Studio:

Your project will serve as a workspace to organize and save your progress.

1. Access Azure AI Studio: Start by navigating to the Home page of Azure AI Studio.
2. Create a New Project: Click on “+ New project.”
3. Name Your Project: Enter a project name.
4. Select or Create a Hub: Choose a hub from the dropdown menu to host your project. If you don’t have one, select “Create a new hub.”
   • Note: To create a hub, you must have Owner or Contributor permissions on the selected resource group. It’s advisable to share the hub with your team to share configurations and centrally manage security settings and spending.
5. Configure Hub Details: If creating a new hub, provide a hub name.
6. Choose a Subscription: Select your Azure subscription from the dropdown. This choice determines billing, access, and administrative settings.
7. Select a Resource Group: Either use the default resource group or choose an existing one from the dropdown.
8. Set Location: Enter the hub’s location, which determines the region for hosting. Azure AI services may have region-specific availability.
9. Select an Azure AI Services Resource: Choose an existing resource or create a new one.
10. Review and Create: On the Review and finish page, check the details and click “Create a project.” The project will be created once the process is complete.

Once your project is set up, you can access the playground, tools, and other assets from the left navigation panel.

Deploying a Chat Model:

To deploy an Azure OpenAI chat model for your copilot:

1. Sign in to Azure AI Studio: Go to the Home page.
2. Access Your Hub: Select your hub from Home > All hubs. If you don’t have one, create it following the instructions.
3. Deploy the Model: From the left pane, go to Deployments > + Deploy model.
4. Choose a Model: On the Select a model page, pick the model you want to deploy (e.g., gpt-35-turbo-16k) and click Confirm.
5. Deploy the Model: Provide a name for the deployment and click Deploy. Once deployed, the deployment details page will show relevant information, such as the creation date and model version.
6. Open the Playground: On the deployment details page, select “Open in playground.”

For more detailed steps, refer to the model deployment documentation.

Chatting in the Playground Without Your Data:

In this quickstart, you’ll test how the model performs without incorporating your data.

1. Access Your Project: Go to your project in Azure AI Studio.
2. Open the Playground: Navigate to Playgrounds > Chat.
3. Select Your Model: Choose your deployed chat model from the Deployment dropdown.
4. Provide a System Message: In the System message text box, enter the prompt: “You are an AI assistant that helps people find information.” You can customize this based on your needs.
5. Optional Safety Messages: Add a safety system message if needed by selecting the Add section button and then Safety system messages. Customize the prebuilt messages to your requirements.
6. Apply Changes: Save your changes and, when prompted, update the system message by selecting Continue.
7. Test the Model: In the chat session pane, ask, “How much do the TrailWalker hiking shoes cost?” and send your question.
8. Review the Response: The assistant may either provide a generic response or state that it doesn’t know the answer, as it lacks access to current product data.

To enhance the model’s performance, you can integrate your data. For more advanced usage, try the Deploy an Enterprise Chat Web App tutorial.