title: 'Get started with Azure OpenAI service'

Azure OpenAI Service brings the generative AI models developed by OpenAI to the Azure platform, enabling you to develop powerful AI solutions that benefit from the security, scalability, and integration of services provided by the Azure cloud platform. In this exercise, you'll learn how to get started with Azure OpenAI by provisioning the service as an Azure resource and using Azure OpenAI Studio to deploy and explore generative AI models.

In scenario for this exercise, you will perform the role of a software developer who has been tasked to implement an AI agent that can use generative AI to help a marketing organization improve its effectiveness at reaching customers and advertising new products. The techniques used in the exercise can be applied to any scenario where an organization wants to use generative AI models to help employees be more effective and productive.

Azure OpenAI service provides a web-based portal named **Azure OpenAI Studio**, that you can use to deploy, manage, and explore models. You'll start your exploration of Azure OpenAI by using Azure OpenAI Studio to deploy a model.

1. In the Azure portal, on the **Overview** page for your Azure OpenAI resource, use the **Go to Azure OpenAI Studio** button to open Azure OpenAI Studio in a new browser tab.

After the new tab opens, you can close any banner notifications for new preview services that are displayed at the top of the Azure OpenAI Studio page.

1. In Azure OpenAI Studio, in the pane on the left, select the **Deployments** page and view your existing model deployments. If you don't already have one, create a new deployment of the **gpt-35-turbo-16k** model with the following settings:

Now that you've deployed a model, you can use it to generate responses based on natural language prompts. The *Chat* playground in Azure OpenAI Studio provides a chatbot interface for GPT 3.5 and higher models.

1. In the **Playground** section, select the **Chat** page. The **Chat** playground page consists of three main panels (which may be arranged right-to-left horizontally, or top-to-bottom vertically depending on your screen resolution):

1. In the **Configuration** panel, ensure that your gpt-35-turbo-16k model deployment is selected.

1. In the **Assistant setup** panel, review the default **System message**, which should be *You are an AI assistant that helps people find information.* The system message is included in prompts submitted to the model, and provides context for the model's responses; setting expectations about how an AI agent based on the model should interact with the user.

1. In the **Chat session** panel, enter the user query `How can I use generative AI to help me market a new product?`

1. Review the response, noting that the model has generated a cohesive natural language answer that is relevant to the query with which it was prompted.

1. Enter the user query `What skills do I need if I want to develop a solution to accomplish this?`.

1. Review the response, noting that the chat session has retained the conversational context (so "this" is interpreted as a generative AI solution for marketing). This contextualization is achieved by including the recent conversation history in each successive prompt submission, so the prompt sent to the model for the second query included the original query and response as well as the new user input.

1. In the **Chat session** panel toolbar, select **Clear chat** and confirm that you want to restart the chat session.

1. Enter the query `Can you help me find resources to learn those skills?` and review the response, which should be a valid natural language answer, but since the previous chat history has been lost, the answer is likely to be about finding generic skilling resources rather than being related to the specific skills needed to build a generative AI marketing solution.

So far, you've engaged in a chat conversation with your model based on the default system message. You can customize the system setup to have more control over the kinds of responses generated by your model.

1. In the **Assistant setup** panel, under **Use a system message template**, select the **Marketing Writing Assistant** template and confirm that you want to update the system message.

1. Review the new system message, which describes how an AI agent should use the model to respond.

1. In the **Chat session** panel, enter the user query `Create an advertisement for a new scrubbing brush`.

1. Review the response, which should include advertising copy for a scrubbing brush. The copy may be quite extensive and creative.

In a real scenario, a marketing professional would likely already know the name of the scrubbing brush product as well as have some ideas about key features that should be highlighted in an advert. To get the most useful results from a generative AI model, users need to design their prompts to include as much pertinent information as possible.

1. Enter the prompt `Revise the advertisement for a scrubbing brush named "Scrubadub 2000", which is made of carbon fiber and reduces cleaning times by half compared to ordinary scrubbing brushes`.

1. Review the response, which should take into account the additional information you provided about the scrubbing brush product.

The response should now be more useful, but to have even more control over the output from the model, you can provide one or more *few-shot* examples on which responses should be based.

1. In the **Setup** panel, under **Examples**, select **Add**. Then type the following message and response in the designated boxes:

**User**:

```