Share via

Manage Interpreter agent for your organization

  • Applies to: Microsoft Teams

APPLIES TO: Image of a checkmark for yes Meetings Image of a x for no Webinars Image of a x for no Town halls

Interpreter agent acts as a translator in Microsoft Teams meetings, allowing participants with a Microsoft 365 Copilot license to listen to the meeting in their chosen language. Interpreter listens to the spoken language in the meeting and translates it into another language in real-time, allowing participants who speak different languages to understand each other and collaborate effectively. To represent their voices, participants can choose to have Interpreter simulate their own voice when translating to others or select one of the following preset automated voices:  Ava (female), Andrew (male), Fable Turbo (neutral). If an automated voice is selected, Ava is the default.

As an admin, you can control whether your organization can use Interpreter and select the default voice option for how users' speech is heard by others.

To learn more about the Interpreter experience in meetings, see Interpreter in Microsoft Teams meetings.

Supported languages

Interpreter supports the following languages for speaking and listening: Chinese (Mandarin), English, French, German, Italian, Japanese, Korean, Portuguese, Spanish.

Prerequisites and licensing

The following list contains the prerequisites for users to access Interpreter in Teams meetings. Users must meet all the following requirements:

Important

Interpreter agent is available as part of the Microsoft 365 Copilot license. A Microsoft M365 license are required to listen to others’ speech through the Interpreter agent.

  • An eligible Microsoft 365 base license.
  • An eligible Microsoft Teams license.
    • Teams licenses might be included in your Microsoft 365 subscription, or you might need to purchase a separate Teams license if you have Microsoft 365 (no Teams) licenses.
  • A Microsoft 365 Copilot license.

Data, security, and privacy

When your users allow Interpreter to simulate their voice, their voice sample isn’t stored.

How Interpreter works

Interpreter performs real-time speech-to-speech (STS) translation using Azure Cognitive Services, supporting multi-speaker, mixed-language conversations in Teams meetings. Here's how it works:

  1. Speech Recognition (ST)- Converts spoken language into English text.
  2. Machine translation (MT)- Translates English text into the selected languages.
  3. Text-to-speech (TTS)- Produces translated speech in the chosen language. TTS can simulate the speaker’s voice or use a predefined voice based on user preference and the admin policy.
  4. A bot transmits meeting audio for cloud-based processing and returns translations instantly.

How Interpreter uses your users' voices

Voice simulation generates translated speech in your user's voice, allowing other participants to hear translations as if they're speaking their language directly. When a user turns on this feature in Interpreter, the system briefly analyzes short segments of their speech on the fly (in real-time) to simulate their unique tone, style, and voice characteristics. Voice samples or biometric data are never stored. AI instantly creates a natural-sounding voice in the selected language, preserving their authentic tone, pitch, and speaking style without exaggerating emotions. This ensures a familiar and seamless multilingual conversation experience.

Real-time processing without storing voice data

Voice data is processed immediately, entirely on the fly, without ever storing your users' voice samples or biometric information. The following diagram illustrates this real-time and secure processing:

  • Original audio streams are briefly analyzed by the system in real-time.
  • ACS Speech services instantly provide translated speech simulation.
  • No voice samples or biometric data are retained after processing.

Architecture diagram of language media processing to ACS speech.

Manage Interpreter using PowerShell

You must use PowerShell to manage Interpreter for your entire organization.

To manage Interpreter for your entire organization, you can use the -AIInterpreter and -VoiceSimulationInInterpreter parameters in the PowerShell CsTeamsMeetingPolicy cmdlet.

Turn Interpreter on or off

The org-wide -AIInterpreter parameter controls whether your users with a Microsoft 365 Copilot license can use Interpreter during meetings in your organization. This parameter is enabled by default.

To turn off Interpreter for your entire organization, use the following script:

Set-CsTeamsMeetingPolicy -Identity <policy name> -AIInterpreter Disabled

To turn on Interpreter for your entire organization, use the following script:

Set-CsTeamsMeetingPolicy -Identity <policy name> -AIInterpreter Enabled

Set the default value for how users can set their preference for how others hear their speech via Interpreter

The org-wide -VoiceSimulationInInterpreter parameter controls your users' default value for Choose your voice when interpreted in Interpreter settings. The Choose your voice when interpreted setting controls how a user's voice is represented to other participants. By default, this parameter is set to enabled.

Here's the user experience for Interpreter depending on the value you choose:

  • Enabled: Sets the default value for Choose your voice when interpreted to Simulate my voice. When users with a Microsoft 365 Copilot licesne turn on Interpreter, it automatically simulates their voices when translating to others in meetings. All participants in an Interpreter-enabled meeting can also select an automated voice. This is the default value.

  • Disabled: Sets the default value for Choose your voice when interpreted to Automated voice. When users with a Microsoft 365 Copilot licesne turn on Interpreter, by deafult, Ava (female) option is selected by default among the automated voices. All participants in an Interpreter-enabled meeting can select another automated voice or choose to have Interpreter simulate their own voice.

To set the org-wide default value for the Choose your voice when interpreted setting to Simulate my voice, use the following script:

Set-CsTeamsMeetingPolicy -Identity <policy name> -VoiceSimulationInInterpreter Enabled

To set the org-wide default value for the Choose your voice when interpreted setting to Automated voice, use the following script:

Set-CsTeamsMeetingPolicy -Identity <policy name> -VoiceSimulationInInterpreter Disabled

Supported platforms and clients

Supported

Interpreter is supported on the following platforms and clients:

  • Teams desktop (Windows and Mac)
  • Teams mobile (iOS and Android)
  • Teams web (Chrome, Microsoft Edge, Safari, and Firefox)
  • Scheduled meetings
  • Channel meetings
  • Virtual Desktop Infrastructure (VDI)

Not supported

Interpreter isn't supported on the following platforms and clients:

  • Unscheduled 1:1 calls (VoIP or Public Switched Telephone Network (PSTN))
  • Meetings scheduled using Microsoft Teams Rooms or personal devices
  • Town halls
  • Webinars
  • Microsoft Teams free