Video Conversation Alerts Using Transcript Relatedness Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current digital communication platforms lack the ability to provide dynamic conversation alerts, such as indicators or alerts triggered in uttered sentences in specific categories like 'budget' or 'intent to buy', and do not allow users to define alert phrases or actions automatically during remote meetings.
Innovation Solution
A system that presents a user interface for users to submit alert phrases associated with categories, processes a communication session transcript to determine relatedness between utterances and alert phrases, and transmits related categories with timestamps for real-time alerts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Difficulty of detecting and measuring
If real-time conversation analysis is implemented to identify alert phrases, then communication analysis capability is enhanced, but system complexity increases
Solution Approach 1:
The patent introduces an intermediary processing system that sits between the communication platform and users. This intermediary automatically analyzes conversation transcripts, identifies alert phrases, and generates notifications, thereby enhancing conversation analysis capability without requiring direct complex processing in end-user devices.
Solution Approach 2:
The system enables self-service by automatically performing conversation analysis and alert generation without manual intervention. The platform autonomously processes transcripts, detects predefined alert phrases, and notifies relevant parties, reducing the need for manual conversation monitoring while enhancing analysis capability.
2Adaptability or versatility
If dynamic alert phrases are allowed during communication sessions, then adaptability improves, but device complexity increases
Solution Approach 1:
The patent implements dynamic alert phrases that can be added, modified, or removed during active communication sessions. Users can customize alert phrases in real-time based on conversation needs, enhancing adaptability. The system dynamically updates the alert phrase database and continues monitoring without requiring session restart or complex reconfiguration.
Solution Approach 2:
The system performs preliminary actions by pre-defining alert phrases and categories before communication sessions begin. Users can set up their preferred alert phrases and categories in advance, allowing the system to be ready for immediate detection and notification when sessions start, reducing real-time processing complexity.
3Productivity
If automatic alert notifications are transmitted in real-time, then productivity increases, but information processing load increases
Solution Approach 1:
The patent extracts and processes only relevant portions of conversation transcripts that match predefined alert phrases. Rather than analyzing entire transcripts in real-time, the system extracts specific segments containing alert phrases and generates notifications only for those relevant portions, reducing overall processing load while maintaining productivity.
Solution Approach 2:
The system implements efficient processing by skipping non-relevant conversation segments and rapidly processing only those portions that contain alert phrases. This selective approach allows real-time notification delivery without requiring continuous full-transcript analysis, reducing processing load while maintaining communication efficiency.
Data Source
AI summary
Dynamic conversation alerts are provided within a communication session. In one embodiment, the system presents, to a client device associated with a user of a communication platform, a user interface (“UI”) including a prompt for the user to submit one or more alert phrases, each alert phrase being associated with a category; receives, from the client device, a list of submitted alert phrases; and receives a transcript of a communication session between participants. For each utterance in the transcript, the system determines whether one or more predictions of relatedness are present between the utterance and one or more alert phrases from the list of submitted alert phrases. The system then transmits, to the client device, a list of related categories, each related category including one or more timestamps of utterances for which a prediction of relatedness is present for an alert phrase associated with that category.