The Challenge
A large regional US bank was struggling with severe customer service and operational issues across two domestic call centers and three supporting vendor locations handling customer service, fraud, and collections calls.
The symptoms were impossible to ignore. Average speed of answer (ASA) sat at 57 seconds, nearly four times its 15-second goal. Post-call customer surveys revealed satisfaction scores averaging 67%, far below the 95% target. Collections and fraud outbound teams were consistently failing to hit goals because they were frequently pulled from their work to answer inbound calls whenever ASA became unmanageable.
The result? Three consecutive years of fraud and collections loss rates exceeding forecasted targets.
The bank had invested in a well-known workforce management (WFM) software package, but the platform couldn’t incorporate the numerous data streams needed to account for account growth and changing call trends. It couldn’t accurately predict future call volumes, optimize staff levels and schedules, or reconcile inbound and outbound call routing in a way that protected both service levels and loss targets.
The PAG Solution
PAG deployed our proprietary enterprise data platform, GOBLIN, to consolidate all relevant data streams into one hierarchical table environment. This gave the bank a single source of truth across systems that had never talked to each other before.
We built a comprehensive call forecast model for both inbound and outbound streams using historical trends, average call times, new account forecasts, delinquent account forecasts, and other critical data to predict call volumes at 30-minute interval levels.
PAG then developed a custom file process to feed the bank’s existing WFM platform with call predictions for the next 90 days on a rolling 12-month basis. This transformed the WFM system from a scheduling tool working with incomplete data into a strategic forecasting engine.
Working closely with the bank’s operations team, PAG rebuilt its scheduling models and realigned FTE to hour intervals to better handle actual call volume patterns rather than assumed demand.
The Results
PAG’s new model – completed in less than three months — revealed something surprising: The bank was understaffed by only 10%, not the 25% they had believed. This single insight changed the entire conversation from emergency hiring to strategic reallocation.
Once a shift bid was conducted to move employees into the new schedules PAG designed, results appeared immediately:
- Month 1: ASA dropped from 57 seconds to 31 seconds with no new hires through better scheduling alignment.
- Months 2-4: After modest hiring to close the actual 10% gap, ASA levels fell under 15 seconds each month, averaging just over 12 seconds.
- Year 1 vendor cost savings: 14% reduction ($2.1 million) as emergency vendor ramp-ups became unnecessary.
- Customer satisfaction: Average scores for the last six months reached 91%, with a peak of 96% in October 2024.
- Loss performance: Combined fraud and collections losses beat 2024 forecasted numbers by $1.7 million as outbound units could finally focus on outbound work instead of covering inbound spikes.
- Strategic reinvestment: The bank used annual savings to invest in new technology and data sources, enabling 14% new account growth in 2025, its highest growth rate in eight years.
- Ongoing engagement: The client has since engaged PAG to help rebuild its internal IT telephony setup to further reduce costs and optimize service levels.
