There are a number of ways to use OpenAI-based services such as ChatGPT in the Azure environment. Depending on your context, you may be willing to use a public service such as ChatGPT for your generative needs. However, if you want to configure OpenAI to work against your own dataset within the boundaries of your Azure tenant, you’ll want to request access to Azure OpenAI Service.

One of the many benefits of enabling OpenAI in your Azure tenant is that you can connect it to your own tenant data. When you start to review the terms of service for public or free AI interfaces, you’ll notice that the service providers can claim ownership over the results that you produce, as well as use the data you supply to further train models. If your goal is to use AI to work against your own proprietary data (which may contain anything from personally identifiable information (PII) to trade secrets), you need to be assured that you remain in control (both from a...

While anyone can browse to OpenAI and use ChatGPT for free interactively, you’ll need to obtain API access if you want to start using ChatGPT in your apps and automations.

To sign up for API access, follow these steps:

1. With a browser, navigate to https://beta.openai.com/signup.
2. On the Create your account page, enter an email address and click Continue. Alternatively, you could use an OAuth connection to authenticate with Google, a Microsoft account, or an Apple ID.
3. Enter a password for your new OpenAI account and click Continue.
4. Check the account you provided during step 2 for a confirmation message. Click the link to confirm your identity and proceed to the onboarding.
5. On the Tell us about you onboarding page, add your personal details, including your Full name, Organization, and Birthday details. Click Agree.
6. If you have registered previously, you may receive a message about upgrading to a paid plan to start...

There are a number of tools that you can use to build and deploy AI-based applications. For purposes of this book, you may need to download and deploy a handful of tools:

• Visual Studio Code:
• Azure Developer CLI:
• Python 3.9+:
• Node.js 14+:
• Git:
• PowerShell 7+: https://github.com/powershell/powershell

Download and install each of those using the provided links. Be sure to use the default installation options for all of them, as any examples in this book will reference default options and paths.

Lastly, we’re going to set up the necessary subscriptions and prerequisite services for Power Platform.

AI models, available through AI Builder in the Power Platform tooling, require AI Builder licensing as well as access to Microsoft Dataverse. To get started, follow these steps:

1. Navigate to the Microsoft Power Automate web portal (https://make.powerautomate.com) and sign in with an administrative account.
2. From the navigation menu, select AI hub.
3. On the Discover an AI capability page, select AI models.
4. If prompted, select Create a database to create a new Dataverse database.
5. On the New database page, select a Currency and Language value, and then click Create my database.
6. If you don’t already have AI Builder capacity, you should see a Start free trial banner at the top of the page, as shown in Figure 2.9. Select it:

Figure 2.9 – Starting a free trial of AI Builder

...

In this chapter, you configured all of the prerequisites for using AI services—including OpenAI in Azure, ChatGPT, and Power Platform AI Builder. You also configured your workstation with the necessary tools to deploy AI apps to Azure.

Next, we’re going to start interacting with ChatGPT as an end user to get an idea of how it works.