Works
How I work
About
Get in touch
interface

Call Quality Control Tool for Americor

Summary

To improve the quality of sales calls and increase the Contact-to-Enrollment ratio, we developed a scalable call quality control system for Americor. Previously, the entire review process was manual: managers listened to random recordings without any structured flagging or issue visualization system. This new tool became the first attempt to systematize call evaluation and automate error detection. We started with deep research into user roles and pain points, which shaped our design approach and prioritization.

Summary
About Americor
Americor is a fintech company specializing in debt resolution services (Debt Settlement and Debt Consolidation Loans). Before this project, the team had no automated system for reviewing calls or managing red flags—everything was done manually. We created a tool to make this process transparent, manageable, and scalable.

We began by analyzing the current workflow and pain points across the sales, compliance, and analytics teams. Key issues included:
Sales Managers
  • No centralized dashboard to track red-flagged calls
  • No prioritization: hard to tell which calls need review
  • Inability to quickly listen to a flagged call and review a transcript
  • Lack of trust in AI-scoring logic
  • Unclear who is responsible for confirming or dismissing flags
  • No historical view of confirmed/dismissed flags
Debt Consultants
  • No access to personal red flags and scoring breakdowns
  • Unclear reasons for low-scoring calls
  • No feedback loop or self-coaching system
Systemic Problems
  • Unconfirmed calls remain in "Unvalidated" status
  • No reports segmented by flag type, team, or channel
  • No customer journey visualization across interactions
  • Fragmented data across Five9, CRM, and Scorecards
  • Compliance team struggles to identify recurring violations

This research laid the foundation for our product hypotheses, architecture, and initial delivery plan.

Product Vision & Hypotheses

Based on the internal presentation, we defined the following vision and assumptions:

Goal:
Create a scalable tool to help managers improve call quality and increase enrollment conversions.
Hypotheses
  • Automatically generated red flags will increase transparency and scalability
  • Visualizing flagged calls will help managers act faster
  • Manual confirmation/dismissal will build trust in AI scoring and close the feedback loop
Personas
  • Sales Director
  • Sales Manager
  • Debt Consultant
  • Compliance Officer
Solution Architecture: Modules & Flows
Initial MVP Interface
In the first version, we created a simple tool that showed a list of calls with basic information, transcript, and AI-scoring. It had no analytics, filtering, or trend visualization. This MVP gave managers a first testable version and helped us gather critical feedback.
Second Version: Analytics and Role-Based Dashboards
In the first version, we created a simple tool that showed a list of calls with basic information, transcript, and AI-scoring. It had no analytics, filtering, or trend visualization. This MVP gave managers a first testable version and helped us gather critical feedback.
Insights & Takeaways

Throughout development, we realized the tool shouldn't feel punitive, but motivational. It should help consultants recognize their weak spots and improve, not feel surveilled.

We documented open product questions: Who should confirm flags—direct managers or any reviewer? How should we distinguish confirmed vs. unconfirmed in the UI? These questions guided UX logic and access permissions.

Next
Design System for Panta Rei