Predictive Dialing Algorithm for Call Center Agent Utilization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Predictive modeling has not been effectively applied to outbound call campaigns due to the complexity of dynamic variables and lack of standard quality training datasets, leading to inefficiencies in call origination and routing in call centers.
Innovation Solution
A computer-implemented method and system that uses a predictive algorithm to determine the number of agents available, create an ordered list of calls, and dynamically dial phone numbers based on probabilities of pick-up, ensuring that the number of calls launched matches the available agents, thereby optimizing call origination and routing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional outbound call campaigns are used without predictive modeling, then the system is simpler to implement, but agent idle time increases and call abandonment rates rise
Solution Approach 1:
The system performs preliminary actions by training predictive models on historical data before actual outbound call campaigns. The models learn patterns from past call outcomes, agent performance, and customer responses, enabling the system to predict future call success probabilities and optimize call pacing in advance rather than reacting in real-time
Solution Approach 2:
The system implements continuous feedback loops where actual call outcomes and agent performance data are fed back into the predictive models. This allows the models to continuously learn and improve their predictions, adjusting call pacing strategies based on real-world results while maintaining system adaptability
2Productivity
If call pacing is increased to maximize agent utilization, then productivity improves, but call abandonment rate increases
Solution Approach 1:
The system dynamically adjusts call pacing based on real-time conditions and predictive model outputs. Rather than using fixed call rates, the system modifies dialing speeds, pause durations, and call distribution patterns according to predicted call success probabilities, agent availability, and historical performance patterns, enabling flexible optimization of both productivity and reliability
Solution Approach 2:
The system changes multiple parameters simultaneously including call interval timing, dialing speed, pause durations between calls, and agent allocation patterns. These parameter adjustments are coordinated based on predictive model outputs to optimize the balance between maximizing agent utilization and maintaining acceptable call completion rates
3Quantity of substance
If more agents are allocated to outbound calls, then call volume capacity increases, but agent idle time increases when calls are not available
Solution Approach 1:
The system enables agents to effectively serve themselves by providing real-time guidance through the predictive model. The system automatically routes calls to appropriate agents based on predicted success probabilities and agent expertise, reducing the need for manual call distribution and minimizing idle time between calls for each agent
4Productivity
If predictive modeling is applied to outbound call campaigns, then call pacing and agent allocation are optimized, but the system complexity and data processing requirements increase
Solution Approach 1:
The system segments the complex predictive modeling task into distinct components: historical data collection and cleaning, feature engineering from multiple data sources, model training on segmented datasets, validation against holdout data, and deployment as separate predictive services. This modular segmentation reduces overall system complexity while maintaining predictive accuracy
Data Source
AI summary
A computer implemented method and computer system for generating outbound calls in a call center in which a plurality of agents are to communicate respectively with outside parties through the outbound calls. The method includes determining a number of agents available for outbound calls, determining phone numbers respectively corresponding to the outside parties, creating an ordered list of calls corresponding to the phone numbers and periodically generating new call attempts by automatically dialing the phone numbers at a dynamic rate, wherein a number of new call attempts is based on a predictive algorithm.