Speech Analytics Workforce Scheduling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Contact centers face challenges in accurately predicting workload and matching staff with appropriate skills to handle varying call volumes and types, leading to inefficiencies and customer dissatisfaction due to overstaffing or understaffing.
Innovation Solution
A system and method that analyzes recorded interactions to classify call reasons, computes agent effectiveness, forecasts demand based on granular interaction reasons, and generates schedules to ensure the right number of skilled agents are available to handle forecasted workloads.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional WFM software is used for forecasting and scheduling, then basic staffing levels can be maintained, but accuracy in predicting workload and matching agent skills to call types is insufficient
Solution Approach 1:
The patent segments the forecasting process by analyzing recorded interactions and classifying them into specific interaction reasons (e.g., billing, technical support, sales). This granular classification enables precise workload prediction for different call types rather than treating all calls uniformly, thereby improving measurement precision without requiring an overly complex system.
Solution Approach 2:
The patent introduces speech analytics as an intermediary component that bridges traditional WFM software and workload forecasting. The speech analytics system processes recorded interactions to extract interaction reasons and agent effectiveness metrics, which then feed into the forecasting and scheduling modules, enhancing accuracy without directly complicating the core WFM system.
2Reliability
If staffing levels are increased to handle peak demand, then customer service availability improves, but unnecessary costs increase due to overstaffing during low-demand periods
Solution Approach 1:
The patent implements dynamic scheduling by continuously analyzing interaction data to compute agent effectiveness metrics and updating forecasts based on actual performance. This allows the system to adapt staffing levels dynamically to match actual workload demands and agent capabilities, ensuring service availability when needed while avoiding overstaffing during low-demand periods.
Solution Approach 2:
The patent changes the parameter of forecasting from static historical averages to dynamic predictions based on computed agent effectiveness metrics. By incorporating real-time performance data and interaction classification, the system adjusts staffing forecasts to reflect current agent capabilities and workload patterns, optimizing the balance between service availability and staffing costs.
3Ease of operation
If agents are scheduled based on general skills only, then scheduling simplicity is maintained, but customer satisfaction decreases due to mismatch between agent capabilities and call complexity
Solution Approach 1:
The patent applies local quality by computing agent effectiveness metrics specific to each interaction reason (e.g., billing, technical support, sales). Instead of using a single general skill rating, the system evaluates agent performance locally for each call type, enabling precise matching of agent capabilities to specific call complexities while maintaining scheduling feasibility through automated computations.
4Measurement precision
If manual analysis of recorded interactions is performed to assess agent performance, then detailed effectiveness metrics can be obtained, but time consumption and computational resources increase significantly
Solution Approach 1:
The patent implements self-service by having the system automatically analyze recorded interactions using speech analytics to compute agent effectiveness metrics. The system independently classifies interactions, identifies interaction reasons, and calculates performance metrics without requiring manual review, thereby achieving detailed effectiveness measurement while minimizing time consumption and computational resource requirements.
Data Source
AI summary
A method for generating an agent work schedule includes: analyzing, on a processor, a plurality of recorded interactions with a plurality of contact center agents to classify the recorded interactions based on a first plurality of interaction reasons and a plurality interaction resolution statuses; analyzing, on the processor, the classified recorded interactions to compute agent effectiveness of an agent of the plurality of agents, wherein the agent effectiveness corresponds to an interaction reason of the first interaction reasons; forecasting, on the processor, a demand of the contact center agents for a first time period for handling interactions classified with the interaction reason; and generating, on the processor, the agent work schedule for the first time period based on the forecasted demand and the computed agent effectiveness.