Overview
Microsoft Teams now supports intelligent speaker attribution in eligible Teams Rooms with intelligent speakers. When enabled, enrolled users are automatically recognized, and meeting transcripts and recordings will identify who is speaking — showing their name instead of generic labels like “Speaker 1” or “Speaker 2.”
This helps improve meeting clarity, accessibility, and searchable transcripts.
Voice recognition and attribution only work in Teams Rooms licensed with a Teams Rooms Pro licence and equipped with an approved Intelligent Speaker device (e.g. Yealink MSpeech, EPOS EXPAND Capture 5).
It does not apply to regular Teams meetings where participants join from their own computers, webcams, or headsets.
Only users with an active voice profile can be identified in meeting transcripts.
How It Works in Meetings
- When you join a meeting in a Teams Room or BYOD room:
- Your voice is recognized using your enrolled profile.
- Your name appears beside your transcripted speech.
- If you haven’t enrolled, your speech will show as “Speaker” or “Speaker 1/2/3” depending on the meeting policy.
Your voice profile works in all meetings where transcription is enabled
Managing Voice Profiles & Opting In/Out
Speaker recognition in Teams Rooms requires users to enrol a voice profile. Users are in full control of their participation:
Opting In – How to Enrol Your Voice Profile
To enable speaker recognition, you must create a voice profile in Teams:
- Open the Teams Desktop App (Windows or macOS).
- Click your profile picture → Settings → Recognition tab.
- Select Enrol voice profile.
- Follow the on-screen prompts and read aloud several short phrases.
- Once complete, your unique voice signature is securely stored in Microsoft 365.
If your voice profile doesn’t appear or your name is missing in meeting transcripts, try deleting and re-enrolling under the same Settings → Recognition page.
Opting Out – Removing Your Profile
Users who prefer not to be identified can simply not enrol a voice profile.
If you’ve already enrolled, you can delete your profile at any time:
Teams → Settings → Recognition → Delete voice profile.
Once deleted, your speech in transcripts will revert to “Speaker”, and Teams will no longer identify you by name.
You can re-enrol your voice profile at any time if you change your mind.
Only users who have created a voice profile can be identified in meeting transcripts. Users who do not enrol or who delete their profile remain anonymous.
Privacy and Security
Your voice data is stored securely in your organization’s Microsoft Cloud tenant and used only to attribute your speech during Teams meetings and recordings. You can delete your voice profile anytime under Settings → Recognition. You can also download your Profile data from the same location.
Supported Languages
Speaker recognition currently supports over 30 languages including English (UK, US, Canada, Australia), French, German, Spanish, Italian, Japanese, and more.
Admin Information
The Global Teams Meeting Policy is now enabled on the tenant with the ‘Distinguish’ setting for Speaker Recognition. This identifies individual speakers as ‘Speaker 1’, ‘Speaker 2’ but DOES NOT apply names to the users within transcripts.
There is also an ‘Attribute’ policy setting for Speaker Recognition – this DOES assign individuals by name within Transcripts.
While the global policy is assigned to all accounts it only applies to Teams Room Pro accounts and devices with relevant supported intelligent speaker hardware.
A summary of the options is below –
🔹 Distinguish
Every speaker is separated and numbered (Speaker 1, Speaker 2, etc.)
Names are never used, even if someone has a voice profile.
Useful if you want speaker separation without personal identification.
🔹 Attribute
Enrolled users (those who have a Teams voice profile) will be identified by name in the transcript.
Non-enrolled users will appear as “Speaker” (not numbered).
This is the only mode that uses actual identity recognition from the voice profile system.
As this is not the default meeting policy setting, Local Admins can raise a request to have their Teams Rooms moved from the global ‘Distinguish’ policy to the ‘Attribute’ policy.
- The Teams Room account is then assigned to the relevant policy
- Any Voice enrolled users will then appear with their names in all future transcriptions, instead of ‘Speaker 1’, ‘Speaker 2’ etc.
Licensing & Device Requirements
- A Teams Rooms Pro license is required for devices to support intelligent speaker recognition.
- Room systems must include certified intelligent speaker hardware (e.g., Yealink MSpeech, EPOS EXPAND Capture 5, etc.).
- The Teams desktop client must be version 1.6.00.12455 or later to access the Recognition settings page.
References
| Last Reviewed Date | 29/01/2026 |
