Speech Analytics for Automated Call Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Fully manual evaluation of agent performance in customer service call centers is time-consuming and prone to user error due to distractions and multitasking, leading to potential missed events during calls.
Innovation Solution
Implementing speech analytics to automatically suggest answers to evaluation form questions based on audio data from calls, reducing the workload and subjectivity of evaluation agents by processing audio data against specified constraints to determine evaluation information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual evaluation of agent performance is used, then evaluation agents can exercise judgment and flexibility, but the process becomes time-consuming and prone to user error
Solution Approach 1:
The patent introduces speech analytics technology as an intermediary between the audio data and the evaluation form. The speech analytics engine automatically processes call recordings, extracts relevant information, and populates evaluation forms, serving as a mediator that reduces manual effort while maintaining evaluation quality
Solution Approach 2:
The system enables self-service evaluation by automatically generating evaluation forms and suggesting answers based on speech analytics. The evaluation process serves itself by using automated data extraction and population, reducing dependency on manual evaluation agent input for routine tasks
2Measurement precision
If evaluation agents manually fill out evaluation forms, then they can capture nuanced performance details, but distractions and multitasking cause them to miss pertinent events
Solution Approach 1:
The patent replaces the mechanical human evaluation process with automated speech analytics technology. The system uses computational algorithms to analyze audio data, detect pertinent events, and extract information, substituting human manual review with automated processing that eliminates distractions and multitasking issues
Solution Approach 2:
The system provides feedback to evaluation agents by automatically suggesting answers based on speech analytics findings. This feedback mechanism allows agents to review automated suggestions and make adjustments if needed, combining automated detection with human judgment
3Productivity
If automatic speech analytics are used to suggest answers, then workload on evaluation agents is reduced, but the system requires processing and analyzing audio data against constraints
Solution Approach 1:
The patent segments the evaluation process into distinct components: speech analytics engine for audio processing, constraint-based answer suggestion for structured evaluation, and automated form population. This segmentation allows each component to specialize in specific tasks, improving overall efficiency while managing complexity through modular design
Data Source
AI summary
A method for determining evaluation information using audio data includes receiving a specification of an evaluation form, the evaluation form including a number of questions, each question having a number of possible answers. A user is received indicating that a first answer of the number of possible answers for a first question of the number of questions should be determined automatically from the audio data, the audio data including speech of a subject. A user input is received associating a first constraint of a plurality of constraints with the first answer. The audio data is processed to determine whether the audio data putatively satisfies the first constraint. The first answer is assigned as an answer to the first question if the audio data putatively satisfies the first constraint. The first answer is stored as the answer to the first question in the evaluation form.