How to Use Microsoft Teams Intelligent Speakers to identify in-room participants in meeting transcription

As a Microsoft Teams member, you have the ability to host meeting in rooms that are configured with Intelligent Speakers. This will allow you to identify participants in real time by checking the live transcripts of the meeting that are transcribed after each meeting. There’s no need to worry about who’s saying what, as all participants can easily see in the meeting what is being said, and the transcripts created after the meeting identify both remote and in-person participants (except for those who do not choose to name themselves).

How it works

An intelligent speaker for Microsoft Teams

The following steps are necessary in order to complete the process.

1. In Teams Rooms, your IT administrator is responsible for setting up Intelligent Speakers. I have written an article that will teach you how to use the keypad controls on an Intelligent Speaker that are based on voice recognition technology.
2. Arrange a meeting with an Intelligent Speaker in a room that has been equipped with one.

   1. Notes: 

      • As a rule of thumb, the number of people on the invite (including yourself) should be no more than 20 in order for Intelligent Speakers to provide voice identification,
      • The number of people in the room should not exceed 10 in order to optimize voice identification.

   It is through the identification of a person’s digital voice profile that the system can be able to determine if a person is attending the meeting in person rather than a remote participant, so that the transcription can be customized in such a way that it represents their voice as accurately as possible. There is only one step involved in this process.

3. Note: If a speaker is not logged into their profile or is not in their same group as the organizer (administrative tenant), then their names will be displayed in red on the transcript, respectively, and their names will be identified as “Speaker 1”, and “Speaker 2”.

4. For the Intelligent Speaker to function optimally, it is recommended that it remain at least 8 inches (20 cm) away from walls and large objects, such as laptop computers, during the meeting so that audio can be clearly heard.
5. Come and join us in the Microsoft Teams Rooms for the meeting.
6. As soon as you are in the meeting, you can start the live transcription in Teams on the desktop.
   • It is possible for presenters who belong to the same tenant as the organizer to prepare a transcription of their speech as soon as their presentation is over.
   • To start and end a meeting, you can speak a command that will allow you to operate the system via voice commands;
7. There will be an opportunity for all attendees at the meeting to edit the transcript during the meeting so that any erroneous identifications will be corrected or participants who were listed as “Speaker X” can be identified, the participants listed incorrectly can be removed, and any participant incorrectly listed will be identified. If a person wishes to remain anonymous, they can choose not to participate in the identification process.
8. The transcript of the meeting can be downloaded after the meeting has taken place.
9. The saved transcript offers attendees the opportunity to manually identify speaker X using the saved transcript, which offers them a second opportunity to correct the identifying information that was used while attending a meeting.

Set up your digital voice profile

For now, setting up your voice profile on Teams has to be done through the desktop version of Teams, which can be used on either Windows or Mac.

1. It would be best to record the recording in a quiet area. It is also recommended that you use the best microphone available to you.
2. It is advised that you set the language for your Teams to English before using it. The English language regions in which you are enrolled can be EN-US, EN-GB, EN-CA, EN-AU, IE (Indian English), or NZE (New Zealand English).
   You can find the language option by selecting More options > Settings, next to your profile picture.

   Note: As soon as your voice profile is set up, you will be able to change it to any of the 17 languages that our system supports.

3. Select Recognition under Settings, then click Get started, and then you will be able to start using it.
4. Click on the Start voice capture button on the next screen, and then read the text that appears on the screen.
5. Select the Stop option when you have finished recording your voice.
6. The final step is to select the Close button on the final screen.

Identify “Speaker X” during live transcription

You are able to attribute the rest of what the other participants in the meeting say to the speaker of the person who made the identification during the meeting. As part of setting up their voice profile, they will also be able to be identified in future meetings.

1. You can identify the speaker of any speech in the list, by going to the Identify speaker option above the speech.
2. In the search box that appears when you enter a name in the field, you will be asked to select a speaker by clicking on the name of the speaker once you have entered the name in the field. There will be an option for users to see in the list the people who receive an invitation to the meeting before the meeting if they received it prior to the meeting if they received it.
3. The checkbox can be selected if, once you are in the meeting, you would like to identify the complete speech which was attributed to “Speaker X” in that meeting or just the part which was attributed to him, in that meeting.

If the transcript includes a pencil icon, this indicates that the name of the person has been manually identified, and that a pencil icon will be displayed next to them.

You will be notified in your Teams Activity that a person you identified has been identified. The person can accept or reject the identification by replying to the notification. A copy of the meeting transcript is most likely to be included in the notification so that the person giving the speech can see the part of their speech that was identified as theirs (and they may need to scroll down a bit in order to locate it).

Correct an identification during live transcription

In order to correct a misidentified speech, you can correct one word or all the words that have been attributed to one individual.

1. In the transcript, locate a speech that is misidentified and select the Edit speaker button.
2. Using the search box on the left side of the screen, you will be able to enter the name of the person and then select their name from the list that appears. The list will include every person who received a meeting invitation any time before the meeting.
3. You can choose to change just this one to make corrections to one incorrect identification, or you can choose to change all of their speeches to make corrections to all their speeches.

If the transcript includes a pencil icon, this indicates that the name of the person has been manually identified, and that a pencil icon will be displayed next to them.

It is possible for the person following a meeting to reject the identification by contacting their Team Activity in their Teams account. A copy of the transcript of the meeting can be included with the notification, along with a highlight of the part of the speech that has been referred to in the notification (they may have to scroll to find it).

A speech attributed to “Speaker X” will be given credit if an edited identification is rejected.

Remove an identification during live transcription

If a speaker’s identification is removed, the speaker is now known as “Speaker X” as it no longer has its identification.

• In the transcript, select the speech that you would like to remove the attribution from and click Remove.

Edit identifications in a saved transcript

Please see Download the saved transcript if you would like to download the transcript after a meeting. It will be listed next to the relevant entry in the downloaded transcript if anyone made any changes to the speaker identification text during the meeting.

Following those instructions, it is possible for you to identify an unidentified speaker, correct a misidentification, or remove your identification from a transcript if you follow the instructions set forth earlier. It is crucial that you save all the transcript files using the same name.

The transcript will notify any individual who has been identified manually during the meeting that their identification has been made. It will also be available to the person in the Teams Activity to reject the identification if they wish.

Update or remove your voice profile

In the event that the Intelligent Speaker has difficulty recognizing your voice after the meeting, you can re-record your voice profile.

In future meetings, your speech will not be identified if you remove your voice profile.

1. Click on your profile picture to reveal the More Options > Settings > Recognition menu in which you’ll be able to choose what to choose.
2. If you would like to rerecord your voice, you should select Update or Remove.

Which languages are supported?

All countries and regions have the following options for enrollment and transcription in meeting for enrollment and in-meeting, respectively. The in-meeting locales that are currently supported by the program are 17 in total.

Enrollment locales 

In the following locations, you will be able to enroll your voice for recognition:

In-meeting transcription locales

As soon as you are enrolled, you will be able to identify voices in meetings and their locations in transcriptions once the meeting has been set to one of the following locations:

Notes

• Tenants based in North America are the only tenants who are currently able to use Intelligent Speakers.
• As of now, all regions are supported with expanded language options.
• Attendees must be invited individually to the meeting and will need to be included on the invitation or forwarded in the invitation.
• You will be able to identify only those people who are in the same tenant as the person who starts the transcription, if you use Intelligent Speakers to identify them.
• A medium-sized room that can hold about 8-10 people is the best size for using Intelligent Speakers.
• If more than 20 people are invited to the meeting in the email invitation, then voice identification is not available.
• As well as accessing your content in the Microsoft 365 cloud, you will also be able to access your voice profile in the cloud as part of your content. Your IT administrator can assist you with accessing your data if you need it.
• We use the voice profile you have provided us with only for the purpose of attribution for your comments during meetings that you have provided us with your consent. Your voice profile will only be used by Microsoft for the purposes that you authorize.
• As long as you have not been invited to an Intelligent Speaker meeting in the past three years, your voice profile will be erased after three years.
• The data you have stored in your audio files can be exported at any time by your IT administrator.
• Your IT administrator can provide you with more information if you are not able to access certain features.