Use the conversation PII playground in Microsoft Foundry

The conversation PII playground in Azure AI Foundry lets you detect and redact personally identifiable information (PII) in conversational text interactively, without writing code. You can submit transcript-style input, configure redaction options, and review detected entities before integrating the feature into your application.

• Stable: 2024-11-01 (GA)
• Preview: 2025-11-15-preview

Prerequisites

• Azure subscription. If you don't have one, you can create one for free.
• Requisite permissions. Make sure the person establishing the account and project is assigned as the Foundry Account Owner role at the subscription level. Alternatively, having either the Contributor or Cognitive Services Contributor role at the subscription scope also meets this requirement. For more information, see Role based access control (RBAC).
• Foundry resource. Create a Foundry resource or see Configure a Foundry resource. Alternatively, you can use a Language resource.
• A Foundry project. For more information, see Create a Foundry project.

Role-based access control (RBAC) requirements

Assign the correct roles to your user principal and project managed identity to access the conversation PII playground. Microsoft recommends using Microsoft Entra ID authentication, which enforces role-based restrictions. Key-based authentication grants full access without role checks and should be avoided in production environments.

• Assign the minimum required roles to both your user principal and project managed identity so they can access Foundry features.

• Verify current assignments using Check access for a user to a single Azure resource.

Navigate to the playground

The active project appears in the upper-left corner. To create a new project:

1. Open the project drop-down menu.

2. Enter a project name or select an existing one.

3. Select Create project.

   

There are two ways to access the conversation PII interface:

• Select the Discover tab from the upper-right navigation bar:

  1. In the search bar, enter Azure and press Enter.
  2. Select Azure-Language-Conversational-PII redaction from the results.
  3. Select Open in Playground.

• Select the Build tab from the upper-right navigation bar:

  1. From the left navigation bar, select Models.
  2. Select the AI services tab.
  3. Select Azure-Language-Conversational-PII redaction to go directly to the playground.

Extract PII from conversations

The extract PII from conversations feature detects and masks personally identifying information in conversational text. This feature is designed to handle the unique structure and context of conversations, such as dialogues or chat logs.

1. On the Playground tab, select Azure Language—Conversational PII redaction from the drop-down menu.

2. Select a sample transcript, use the paperclip icon to upload your transcript, or enter conversation turn text.

3. Format your conversation with each turn on a new line and include speaker labels if possible. Inconsistent use of participant ID, colon, message, or new line may produce unexpected output. Here's an example of well-formatted conversation text:

   Speaker 1: Hello, how are you?
   Speaker 2: I'm good, thank you! How about you?
   Speaker 1: I'm doing well, thanks for asking.
   

   Tip

   You can use built-in formatting assistant (Format for me) in the playground to ensure your input meets expected conversation structure.

4. In the Configure pane, set the following options:

   Option	Description
   API version	Select the API version that you prefer to use.
   Model version	Select the model version that you prefer to use.
   Language	Select the language in which your source text is written.
   Select types to include	Select the PII types you want to redact.
   Specify redaction character	Choose the character used to mask sensitive text.

5. Select Detect. Detected entities are highlighted in the text, and you can review the accompanying details in formatted text or as a JSON response.

   

   Field	Description
   Type	The detected entity type.
   Confidence	The model's level of certainty that it correctly identified the entity type.
   Offset	The number of characters from the beginning of the text to the entity.
   Length	The character length of the entity.

Verify that the detected entities match the PII in your input conversation. Select Edit to modify the Configure settings and rerun detection as needed.

Open in Visual Studio Code 🆕

After validating your scenario in the playground, select Open in VS Code to carry your current configuration directly into a development environment—no manual setup required.

1. Configure your scenario in the playground:
   • Select your API version and model version.
   • Enter and test your sample input.
   • Adjust options such as entity types, redaction character, and language.
2. Select Open in VS Code.
3. Visual Studio Code opens with a preconfigured code sample that reflects your playground configuration, including your API version, model, and redaction settings.

Next steps

After opening in Visual Studio Code:

1. Add your authentication credentials.
2. Replace the sample input with your production data.
3. Run and validate the request.
4. Integrate the code into your application workflow.