Enterprise Voice Assistant Control for Multi-Participant Calls
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital assistants are not suitable for enterprise environments due to issues such as multiple users speaking simultaneously and the inability to verify commands, making them ineffective in large offices or business settings.
Innovation Solution
A method for controlling real-time conversations and communication platforms using a digital assistant that identifies an active speaker, verifies authorization, and establishes a dedicated audio channel for improved speech recognition, utilizing components like ASR, NLU, and a virtual AI endpoint to execute commands.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a central digital assistant speech box is deployed in an enterprise environment with multiple users, then voice control functionality is provided, but the system cannot verify commands when multiple users speak simultaneously
Solution Approach 1:
The patent segments the audio processing by identifying and separating individual speakers in the conversation. The system divides the mixed audio stream into distinct speaker channels, allowing independent analysis and verification of each speaker's commands, thus resolving the reliability issue when multiple users speak simultaneously.
Solution Approach 2:
The patent introduces an intermediary speaker identification and verification mechanism between the microphones and the digital assistant. This intermediary layer analyzes audio signals to determine which participant is speaking, verifies their authorization, and routes commands appropriately, enabling reliable command verification in multi-user environments.
2Ease of operation
If digital assistant is activated for every user in a large office, then each user can access voice control, but the system complexity and resource requirements increase significantly
Solution Approach 1:
The patent implements a universal digital assistant system that serves all users through a single integrated platform. Rather than deploying separate digital assistants for each user, the system provides multi-functional support for multiple participants, reducing system complexity while maintaining ease of operation for all users.
Solution Approach 2:
The system automatically identifies speakers and manages authorization without requiring manual configuration for each user. The digital assistant self-adjusts to the current conversation context, dynamically determining which participant is speaking and whether they are authorized to issue commands, thereby simplifying operation while maintaining security.
3Reliability
If audio signals from all participants are continuously analyzed for commands, then no commands are missed, but processing resources are wasted when no one is speaking or when background noise is present
Solution Approach 1:
The patent implements periodic speaker detection and activation triggers rather than continuous command analysis. The system periodically monitors audio signals to detect when a participant is actively speaking and only then activates full command analysis, reducing energy consumption while ensuring commands are not missed during active speech periods.
Solution Approach 2:
The patent extracts and processes only the relevant audio signals from active speakers, separating them from background noise and inactive participants. By taking out only the necessary audio processing tasks (active speaker identification and command analysis) from the continuous processing stream, the system reduces energy consumption while maintaining reliable command detection.
Data Source
AI summary
The present disclosure relates to a method for controlling a real-time communication between at least two participants on a real-time conversation and collaboration platform by means of a digital assistant unit, wherein clients are connected to a conferencing application via a communications network establishing the communication, the method comprising the steps of identifying, from the at least two participants, a first participant as an active speaker in the conversation by using audio signals received from the first participant via a microphone, activating the digital assistant unit for the first participant, if a predetermined event is detected, wherein the audio signals received from the first participant are analyzed so as to identify voice commands therefrom, wherein an Automatic Speech Recognition engine performs a voice recognition procedure for identifying and transcribing the identified voice commands, and wherein the transcribed voice commands are analyzed and executed.