To enable Intelligent Speaker functionality (which allows for speaker identification in a shared room), three distinct layers must be addressed:

1. User Enrollment: Each individual must first create a "face and voice" (digital voice) profile in their desktop Teams app settings so the system has a reference for their voice.
2. Policy Enablement: The administrator must ensure the Meeting Policy has AllowTranscription and VoiceId settings enabled and assigned to the users.
3. Active Execution: During the meeting, Transcription must be manually started (or set to auto-start) for the Intelligent Speaker hardware to begin mapping the captured audio to the digital profiles of the participants.

Microsoft Learn: Set up Microsoft Teams Intelligent Speakers. This document outlines the requirement for users to create a voice profile in the Teams Desktop app under Settings > Recognition.

Microsoft Learn: Manage meeting policies for transcription and captions. Detailed instructions on using the Set-CsTeamsMeetingPolicy cmdlet with the -AllowTranscription parameter to enable identification.

Microsoft Support: Use Intelligent Speakers to identify in-room participants in a meeting transcript. Confirms that transcription must be active for identity recognition to function.

Microsoft Tech Community: Introducing Intelligent Speakers for Microsoft Teams Rooms. Section on "Tenant Admin Controls" for assigning policies to specific users or rooms.