Multi-pass Speech Analytics for Call Center Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Businesses face challenges in efficiently processing and analyzing large volumes of communication data from call centers due to resource-intensive speech recognition technologies, which hinder timely review and analysis of customer service interactions.
Innovation Solution
Implementing a multi-pass analytics system that includes first and second tier speech to text analytics, where the first tier performs expedited speech recognition and sends relevant data to a second tier for more thorough analysis, optimizing resource usage and improving processing speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If speech recognition technology is used to capture and analyze communication data, then the ability to review and analyze customer service interactions is improved, but hardware resources are consumed extensively
Solution Approach 1:
The patent segments speech recognition processing into multiple tiers: a first tier that performs expedited speech-to-text conversion on all audio data, and a second tier that performs more thorough analytics only on selected portions. This segmentation allows the system to maintain high speech analysis capability while reducing overall hardware resource consumption by applying intensive processing only where necessary.
Solution Approach 2:
The patent applies partial action by performing comprehensive speech analytics on only a portion of the total communication data rather than all data. The first tier processes all data at a basic level, while the second tier performs intensive analytics selectively, thus achieving effective speech analysis without the excessive resource consumption of full-depth processing on all data.
2Measurement precision
If thorough speech analytics are performed on all communication data, then analysis accuracy is improved, but processing speed decreases
Solution Approach 1:
The patent divides analytics processing into two segments: a first tier that performs expedited speech-to-text conversion on all audio data to maintain processing speed, and a second tier that performs thorough analytics only on selected portions to maintain accuracy. This segmentation allows the system to achieve both speed and accuracy by applying different processing depths to different data portions.
Solution Approach 2:
The patent applies partial action by performing comprehensive analytics only on a portion of the data rather than all data. The selective application of thorough analytics in the second tier maintains accuracy for critical cases while the expedited first tier maintains overall processing speed across the entire dataset.
3Measurement precision
If all communication data is reviewed, then customer service quality assessment is improved, but the time required for review increases
Solution Approach 1:
The patent segments the review process into two tiers: a first tier that rapidly processes all communication data to convert speech to text, and a second tier that performs detailed quality analytics only on selected portions. This segmentation enables comprehensive quality assessment across all data while maintaining rapid processing by limiting intensive analysis to specific cases.
Solution Approach 2:
The patent applies partial action by performing detailed quality analytics on only a portion of the communication data rather than all data. The expedited first tier ensures all data is processed quickly for basic conversion, while the selective second tier application of thorough analytics maintains assessment accuracy for critical cases without increasing overall review time.
Data Source
AI summary
Included are embodiments for multi-pass analytics. At least one embodiment of a method includes receiving data associated with a communication, performing first tier analytics on the received data, and performing second tier analytics on the received data, where the second tier analytics determines different characteristics of the received data than the first tier analytics.