Adaptive Estimated Wait Time Predictor for Contact Centers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Contact centers face challenges in providing accurate estimated wait times to callers, as existing predictors often require significant resources and are inflexible, leading to frustration and abandoned calls due to lack of clear wait time information.
Innovation Solution
An adaptive Estimated Wait Time (EWT) predictor system that selects from multiple algorithms based on current contact center conditions, such as resource availability and accuracy thresholds, to provide the most suitable EWT predictions, balancing accuracy with resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If complex wait time prediction algorithms are used, then prediction accuracy is improved, but resource consumption increases
Solution Approach 1:
The system dynamically selects from multiple algorithms of varying complexity based on current contact center conditions, queue characteristics, and performance requirements. This allows the system to adapt its computational approach in real-time, using more complex algorithms when high accuracy is needed and simpler algorithms when resource efficiency is prioritized.
Solution Approach 2:
The system changes algorithmic parameters and complexity levels based on varying operational conditions. By adjusting which algorithm is applied based on queue depth, agent availability, and time of day, the system optimizes the balance between prediction accuracy and resource consumption for different scenarios.
2Device complexity
If a single algorithm is used for all scenarios, then system simplicity is maintained, but adaptability to different conditions deteriorates
Solution Approach 1:
The system implements a universal algorithm selection framework that can accommodate multiple different prediction algorithms within a single system. This multi-functional approach allows the same system structure to handle diverse algorithmic requirements while maintaining overall simplicity through a standardized selection and execution mechanism.
Solution Approach 2:
The system segments the algorithm selection process into distinct components: a selection module that evaluates current conditions, a ranking module that orders algorithms by suitability, and an execution module that applies the selected algorithm. This segmentation maintains system simplicity while enabling flexible adaptability to different scenarios.
3Reliability
If multiple algorithms are evaluated and ranked, then optimal algorithm selection is improved, but processing overhead increases
Solution Approach 1:
The system performs preliminary actions by pre-ranking algorithms based on historical performance data and theoretical efficiency metrics before actual runtime selection. This preliminary ranking reduces the computational overhead during real-time operation, as the system only needs to evaluate current conditions against a pre-established hierarchy rather than fully evaluating all algorithms from scratch.
Solution Approach 2:
The system implements feedback mechanisms that use actual prediction accuracy and resource consumption data to continuously refine algorithm rankings. This feedback loop improves selection quality over time while the system learns to minimize processing overhead by identifying which algorithms perform best under specific conditions without requiring exhaustive evaluation each time.
Data Source
AI summary
Systems and methods for providing adaptive estimated wait time predictions for work items are provided. More particularly, a processor-enabled module may select a best estimated wait time algorithm from among a plurality of estimated wait time algorithms that meets the accuracy and calculation resource conditions of a contact center based on rules. Furthermore, the conditions of a contact center may change at times and as a result the selection of which estimated wait time algorithm is considered best will adapt to suit the contact center change. In addition, the selected best estimated wait time prediction can be provided to client communication devices.