A Tier-Based Framework for Matching Agent Capabilities to Inquiry Complexity
Open Access DepositedOPTIMIZING MUNICIPAL CALL CENTER WORKFORCE ALLOCATION
Civil service regulations restrict the ways in which managers can redistributeemployees, but all callers, including an individual who inquires about trash collection and a citizen seeking help with a building code violation, should receive the same quality of service. This strain is particularly high at NYC 311, whose call mix cuts across thousands of inquiry types. This praxis explores whether a system of categorizing call inquiries based on the complexity of their operations, unsupervised machine learning on patterns of resolution and escalation, can optimize the allocation of the workforce in the municipality context with limited resources, and uses NYC 311 as a case study. The research aims at answering four questions
Can 311 questions be classified into operationally distinct levels of complexity reliably? Are the existing staffing levels representative of those levels, or is there a structural mismatch? And does tiered forecasting improve predictions to the extent to realize actual capacity improvements? The three operationally distinct tiers were identified with unsupervised clustering (Silhouette=0.897), and the stablility was assessed using bootstrap resampling which is above the limit of 0.70. Chi-square analysis verified that the mismatch in workforce and demand was systematic (234.14, p<0.001, Cramér's V =0.914), as Tier 1 was overstaffed by 42.6% and Tier 2 was understaffed by 39.9%. Tier-based forecasting had a 25.4% average MAPE reduction, which allowed the capacity to meet 224,800 additional citizens each year without budget enhancements. Results also show that evidence-based workforce assignment can currently be accomplished within civil service limits, providing city managers with a tangible roadmap in matching staff capacity to real-call demand - without recruitment and without workaround regulations.
- All rights reserved
Notice to Authors
If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.