Automated Caller Satisfaction Detection via Input Count Discrepancy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automated calling systems, such as IVR systems, face challenges in determining user satisfaction efficiently, as manual methods are expensive and time-consuming, and current automated methods are not practical for high call volumes, leading to frustration when speech input is not accurately understood.
Innovation Solution
An automated method that maintains actual and ideal input counts during a call, associating no difference between counts with maximum satisfaction and increasing differences with decreasing satisfaction, allowing for adjustments such as changing dialog paths, speech recognition grammar, or routing to a live agent when thresholds are met.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual methods (human listeners) are used to determine user satisfaction, then measurement precision is improved, but productivity deteriorates due to high cost and time consumption
Solution Approach 1:
The system automatically monitors and evaluates its own performance by tracking speech recognition success rates and call progression metrics without requiring external human evaluation, enabling the system to self-assess user satisfaction
Solution Approach 2:
The patent replaces the mechanical system of human listeners with an automated computational system that uses software-based metrics (speech recognition success rates, dialog path progression) to evaluate user satisfaction, eliminating the need for manual call review
2Productivity
If automated systems are used to handle large volumes of calls, then productivity is improved, but user satisfaction deteriorates due to frustration when speech input is not accurately understood
Solution Approach 1:
The system continuously monitors speech recognition success rates and compares actual dialog path progression against expected paths, using this feedback to detect when users are becoming frustrated and to trigger appropriate corrective actions
Solution Approach 2:
The system dynamically adjusts its behavior based on real-time metrics by automatically routing calls to live agents when speech recognition fails repeatedly or when users deviate from expected dialog paths, adapting to changing user needs during the call
3Ease of operation
If speech recognition is used to automate call handling, then ease of operation is improved, but device complexity increases due to the need for accurate speech understanding
Solution Approach 1:
The system segments the call handling process into distinct phases (speech recognition, dialog progression, satisfaction monitoring) and applies different evaluation metrics to each segment, allowing complex speech recognition to be managed through modular, independent monitoring components
Data Source
AI summary
An automated method for determining user satisfaction with a call comprising (1) maintaining a count of the number of times a user was asked for input during the call (the “actual count”), (2) maintaining a count of the number of times a user should have been asked for input based on the dialog path the user traversed (the “ideal count”), and (3) associating no difference between the actual count and the ideal count with maximum user satisfaction and associating increasing differences between the counts with decreasing user satisfaction. In one embodiment, if the difference between the actual count and ideal count reaches a certain threshold, the call system changes the interaction with the user.