IVR Contextual Prediction Selection for Response Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing Interactive Voice Response (IVR) systems rely solely on classical prediction processing, which often fails to provide accurate responses due to a lack of consideration for contextual information from previous user inputs, leading to user frustration.
Innovation Solution
A three-tiered selection method that dynamically chooses between contextual and classical prediction processing based on confidence levels and sentiment analysis to ensure accurate responses, utilizing contextual information when necessary to enhance user interaction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If contextual prediction processing is used to improve response accuracy, then response accuracy is improved, but resource consumption increases
Solution Approach 1:
The system dynamically adjusts the prediction processing mode (classical vs contextual) based on real-time confidence level assessments. When classical prediction achieves sufficient confidence thresholds, the system remains in low-resource mode; when confidence is insufficient, it transitions to contextual prediction mode to improve accuracy, thereby adaptively balancing resource consumption and response accuracy
Solution Approach 2:
The system changes the processing parameters by switching between two distinct prediction modes with different resource requirements. Classical prediction uses fewer resources while contextual prediction uses more resources but provides better accuracy. The system selects the appropriate parameter set (prediction mode) based on confidence level thresholds, effectively managing the trade-off between resource consumption and accuracy
2Measurement precision
If contextual prediction processing is always used, then response accuracy is improved, but system complexity increases
Solution Approach 1:
The prediction system is segmented into two distinct processing pathways: classical prediction processing and contextual prediction processing. Each pathway has its own confidence level thresholds and processing rules. This segmentation allows the system to manage complexity by handling different types of queries through appropriate pathways rather than using a single complex contextual processing system for all cases
Solution Approach 2:
The confidence level assessment mechanism acts as an intermediary that determines whether to invoke contextual prediction processing. This intermediary layer simplifies the overall system by providing a clear decision criterion (confidence thresholds) that controls when the more complex contextual processing should be activated, preventing unnecessary complexity for simple queries
3Use of energy by moving object
If classical prediction processing is used, then resource consumption is reduced, but response accuracy deteriorates
Solution Approach 1:
The system applies partial contextual processing by first attempting classical prediction (partial action) and only invoking full contextual prediction processing when necessary (excessive action only when needed). This approach uses resource-efficient classical processing for most cases while reserving contextual processing for difficult cases where accuracy is compromised, achieving a balance between resource consumption and accuracy
4Ease of operation
If contextual prediction processing is used, then user satisfaction is improved, but processing time increases
Solution Approach 1:
The system uses periodic confidence level assessments at predetermined intervals (first tier, second tier, third tier) to determine whether contextual prediction processing is needed. This periodic checking mechanism allows the system to mostly operate in fast classical mode while periodically evaluating whether a switch to slower contextual processing is necessary, thereby minimizing average processing time while maintaining user satisfaction when needed
Data Source
AI summary
Apparatus, methods and systems for contextual prediction processing is provided. Methods may include receiving a conversation from an entity. The conversation may include current utterance, previous utterances and details. Methods may include using an action-topic ontology to build, using data retrieved from the current utterance, a conversation frame that corresponds to the current utterance. Methods may include merging the conversation frame with data, retrieved from the previous utterances and the details, to generate a target conversation frame. Methods may include validating the target conversation frame to prevent looping over historic data in the event that the current utterance fails to add relevant information. Methods may include generating an enhanced contextual utterance based on algorithms and the target conversation frame. The enhanced contextual utterance may be used to understand the current utterance in a context of the conversation. Methods may include returning the enhanced contextual utterance to the entity.