top of page

THE CHALLENGE

rapid! hired me specifically because they were not receiving adequate support from Green Dot’s internal L&D and customer experience teams. Their existing systems were outdated and inefficient. These poor service metrics directly led to client losses, driven by employee complaints about the subpar user experience.

THE SOLUTION

New CX Tech Stack and L&D

Company background
Green Dot operates seven highly differentiated financial lines of business, including Apple Cash, Walmart PayCards, tax services, and Rapid, making its portfolio diverse but scattered.

rapid! background
A 20-year leader in PayCards, had recently launched earned wage access features and a white-label product to expand its offerings. While the business boasted a strong sales and enrollment team focused on acquiring employers, there was no internal representation for the end users—the employees relying on Rapid PayCards and earned wage access tools

 

​​Approach - Phase 1: Discovery and Findings

  • Outdated Knowledge Base

    • The system (RoboHelp) took 2 days to update due to load times, making real-time updates impossible and reducing agent accuracy.

  • Inefficient Telephony System

    • The UJET telephony system only recorded 60% of calls, making QA reviews and complaint research nearly impossible.

  • Disconnected Tech Stack

    • Metadata from the IVR did not connect to the CRM, leaving agents unprepared and unaware of customer issues during interactions.

  • High BPO Agent Attrition

    • 95% of agents at the BPO did not last beyond 4 months, creating high turnover and constant retraining needs.

  • Ineffective QA Process

    • The QA scorecard was outdated, with poor follow-through—responses were limited to supervisor huddles that lacked actionable insights or measurable improvements.

  • Lack of Actionable Data

    • No linking between call dispositions and CRM data. Reports on call reasons were unavailable, leaving customer experience reactive and data visibility close to zero.

  • Poor Feedback Systems

    • IVR-based CSAT surveys at the end of calls were unreliable, with 40% of calls unrecorded, skewing customer experience metrics.​

Key Metrics when I started
 

  • CSAT: 48% (worst LOB)

  • FCR: 53%

  • Average Call Time: Over 400 seconds.

  • Average Hold Time: Above 5 minutes.

  • NPS: Negative.

Phase 2: Strategy and Deliverables

  • Knowledge Base Migration and Revamp

    • Secured buy-in for KMS Lighthouse, revamping 600+ articles with scenario-based content and enabling integration with Salesforce across six lines of business, improving agent efficiency and knowledge accuracy

  • Telephony System Overhaul

    • Adopted Twilio, ensuring 100% call recording, connecting IVR metadata to agent tools, and automating relevant knowledge base article retrieval, drastically improving agent preparation and QA processes.

  • Improved Tech Stack Integration

    • Integrated Twilio and KMS Lighthouse to share metadata, streamlining workflows for agents, though CRM integration remained limited due to outdated infrastructure.

  • Reduced BPO Attrition

    • Extended new hire training to 12 days, tracked metrics for each class, and used KMS Lighthouse for agent feedback, reducing attrition and increasing engagement.

  • QA Process Overhaul

    • Updated QA scorecards, added micro-trainings for specific failures, and held biweekly calibration meetings, driving consistent scoring and measurable agent improvement.

  • Actionable Data Development

    • Bootstrapped solutions with MicroStrategy to link call dispositions and CSAT scores, providing deeper insights into agent tenure, training impacts, and performance.

  • Enhanced Feedback Systems

    • Implemented Medallia surveys via email, capturing CSAT, NPS, and open-ended responses to inform key decisions and significantly improve customer experience insights.

  • Completely revamped the knowledge base into a modern, 21st-century system with A-B testing for all articles.
     

  • Added feedback and agent rating features to address underperforming articles through direct agent insights.
     

  • Enabled breaking news updates for system issues and product launches.
     

  • Integrated metadata from the IVR to auto-pull the correct article based on the customer’s IVR inputs.
     

  • Implemented scenario-based articles with decision trees to guide agents step-by-step, improving efficiency and accuracy.

bottom of page