Closed Loop Decisionmaking for Automated Care Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Interactive voice response (IVR) systems face inefficiencies due to imperfect automatic speech recognition and natural language understanding, leading to customer frustration, increased costs, and manual update challenges, particularly when supplementing Self Care Systems (SCSs) with Assisted Care Systems (ACSs, which can result in improper transfers and lack of automated software updates.
Innovation Solution
A computer-based system that processes customer interactions using models for context maintenance, data record creation, and automatic model updates, incorporating statistical prediction and semantic classification models to improve IVR system performance, including a decision agent and feedback system for customer satisfaction analysis and model refinement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If automatic speech recognition and natural language understanding software are used to process customer interactions, then the system can understand and respond to natural language statements, but word recognition and context recognition are not perfect leading to repeated prompting and customer frustration
Solution Approach 1:
The system collects feedback data from customer interactions including successful and unsuccessful recognition cases, then uses this feedback to automatically update and refine the speech recognition and natural language understanding models, continuously improving accuracy while maintaining ease of natural language operation
Solution Approach 2:
The system performs preliminary actions by pre-processing and analyzing customer statements before final routing decisions, using statistical prediction models to anticipate customer needs and prepare appropriate responses, reducing the need for repeated prompting
2Loss of information
If the system repeatedly prompts customers for more information due to imperfect recognition, then it attempts to gather complete information, but this causes customer frustration and increased abandoned calls
Solution Approach 1:
The system applies partial action by making routing decisions with incomplete information when confidence thresholds are met, using statistical prediction to estimate missing information rather than requiring complete data, thus reducing repeated prompts while maintaining adequate service quality
Solution Approach 2:
The system uses feedback from interaction outcomes to learn which information gaps are critical and which can be tolerated, automatically adjusting its information gathering strategy to minimize customer frustration while maintaining necessary information completeness
3Adaptability or versatility
If the Self Care System is supplemented with an Assisted Care System with multiple agent groups, then it can handle complex customer requests, but the system may transfer customers to improperly trained agent groups increasing costs and frustration
Solution Approach 1:
The system uses feedback from transfer outcomes and agent performance data to automatically refine its routing models, learning which customer request types should be directed to which agent groups, thereby reducing improper transfers while maintaining the ability to handle complex requests across multiple specialized groups
Solution Approach 2:
The system implements dynamic routing that adapts to changing conditions by continuously updating its understanding of agent capabilities and customer needs, allowing flexible allocation of complex requests to appropriate agent groups based on real-time performance feedback rather than static routing rules
4Reliability
If the system uses manual update processes to improve software performance, then updates can be implemented, but the process is cumbersome and time-consuming
Solution Approach 1:
The system performs self-service by automatically collecting interaction data, analyzing performance metrics, generating model updates, and implementing improvements without human intervention, thus maintaining high software reliability while eliminating the time-consuming manual update process
Solution Approach 2:
The system uses continuous feedback from customer interactions to automatically trigger and guide the software update process, identifying performance issues and implementing corrections autonomously, replacing cumbersome manual updates with an automated feedback-driven improvement cycle
Data Source
AI summary
There is disclosed a system functional for operating and updating an interactive voice response self care system which utilizes various types of models. The system is capable of improving the performance of the self care system through updating the models used in that system in a closed loop manner. The system is also potentially configured to utilize input from a human agent, such as a customer service representative, to which a customer call is transferred by a self care system.