Speech Analytics for Voicemail Prioritization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current voice messaging services require subscribers to listen to each call in order and manually review messages to understand their content and importance, and messages can only be queried by caller name, phone number, or manually annotated keywords, making it inefficient to manage multiple calls and prioritize messages.
Innovation Solution
Implementing speech analytics in a messaging management module to automatically identify message categories, associate language models, and perform messaging management functions, such as prioritizing messages and triggering actions, based on predefined rules, allowing for out-of-order retrieval and automated actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If subscribers manually review each message to understand content and importance, then message content understanding is achieved, but time consumption increases significantly
Solution Approach 1:
The system performs preliminary analysis of message content by automatically transcribing speech to text and extracting keywords before the subscriber needs to review the message. This advance processing enables quick scanning and prioritization without requiring the subscriber to listen to entire messages, thus understanding message importance in advance while minimizing time investment.
Solution Approach 2:
The patent introduces an intermediary processing layer between message reception and subscriber review. This layer includes automatic transcription services and keyword extraction algorithms that convert audio messages into searchable text with identified key terms. The intermediary processes the raw message data into a format that enables rapid scanning and understanding without requiring complete message playback.
2Stability of the object's composition
If messages are retrieved in order received, then chronological organization is maintained, but important messages may be delayed
Solution Approach 1:
The system implements dynamic message prioritization where the retrieval order is not fixed but adapts based on analyzed content characteristics. Messages are assigned priority levels through automatic keyword analysis and categorization, allowing the system to dynamically reorder message queues. High-priority messages identified through content analysis are surfaced to the top regardless of their chronological position, while maintaining the ability to switch between different sorting modes.
Solution Approach 2:
The patent changes the parameter used for message sorting from purely chronological (time received) to a composite parameter that includes content-based priority indicators. The system analyzes message content to extract keywords and determine importance, then uses this analysis to assign priority values that modify the retrieval order. This parameter transformation enables important messages to be accessed faster while subscribers can still view messages chronologically if desired.
3Ease of operation
If messages are queried only by caller name or phone number, then simple lookup is enabled, but content-based searching is limited
Solution Approach 1:
The system implements a universal search mechanism that handles multiple query types through a single interface. The search function can process queries by caller identification (name or phone number), by extracted keywords from message content, by message category, or by combinations of these criteria. The underlying transcription and keyword extraction infrastructure enables all these search modes to operate from the same processed message data, providing versatile searching without requiring separate systems for each query type.
Data Source
AI summary
Embodiments are provided for utilizing speech analytics for automatically managing messages in a communications network. A message from a calling party to an unavailable called party is recorded in a messaging management module. The messaging management module automatically identifies a category for assigning to the content of the message based on information associated with the unavailable called party or a calling party and associates a language model with the message content based on the category. An analysis module utilizes the language model to identify grammar making up the message content. The messaging management module automatically performs a messaging management function on the message content based on a set of rules. The messaging management function may include automatically determining a priority for the message (enabling the message to be retrieved out of order) or automatically triggering an action in response to the message which is directed to the unavailable called party.