Deloitte
2022
Compass: Quality Assurance Platform that Improved Sales Call Evaluation Efficiency by 90%
Designing an AI-powered enterprise platform to automate sales call compliance, mitigate legal risk, and scale review volume 11x for a major U.S. credit card provider
Overview
Scaling AI-Driven Compliance Monitoring for Enterprise Sales
Designed the industry’s first NLP (Natural Language Processing)-powered Quality Assurance platform to help a leading U.S. credit card provider mitigate legal risks, satisfy SEC(Securities & Exchange Commision)/CFPB (Consumer Financial Protection Bureau) regulations, and transform manual review workflows.
The Impact
Slashed individual call evaluation time by 90% (from 60 minutes down to 5 minutes) and scaled weekly evaluation volume by 11x without expanding the QA team - all while securing full regulatory compliance.
Problem
Untangling a 60-Minute Manual Evaluation Bottleneck
The previous QA process was painfully slow, allowing the team to evaluate only 10% of U.S. sales calls. To meet SEC and CFPB regulations, that volume had to jump to 50%. With analysts spending an entire hour manually auditing just one call, scaling headcount wasn't an option - the entire workflow needed a structural overhaul.
400/4000
Total sales calls evaluated versus total sales calls made in a week
10
Number of QA Analysts
1 Hour
Time required by 1 QA Analyst to evaluate 1 Sales Call
Why does it take 1 hour to evaluate a Sales Call?
A QA Analyst manually listens to the entire sales call to identify non-compliant sales practices, which can occur at any point during the call.
Research
Uncovering Analyst Pain Points Through Direct Discovery
Interviewed 7 QA Analysts to map their daily friction. The core insight was clear: Analysts didn't want to listen to entire calls; they wanted the system to surface the right information instantly. As one analyst put it: "I’d love to have a new system that provides me with all the information that I need to monitor a call in one place."
Discovered Analyst Desires
“As a QA Analyst, I need a call preview before evaluation to plan effectively.”
“As a QA Analyst, I need unified guidelines for each product for fair evaluation.”
“As a QA Analyst, I want to prioritize tracking compliance instead of multi-tasking to boost productivity.”
Product Strategy & Direction
Key Insight
Analysts needed to transition from hunting for infractions to reviewing pre-flagged insights
This insight drove our core product strategy: leverage NLP to automate the heavy lifting, eliminating ambiguity through unified data presentation, smart contextual guidelines, and optimized evaluation flows.
Key UX Considerations
User Research Insights led to the prioritization of 3 key solutions that support all QA functions and scale effectively
Analyst Quote | Pain Point | Feature/Solution |
|---|---|---|
“As a QA Analyst, I need a call preview before evaluation to plan effectively” | Lack of call content visibility before evaluation hinders preparation | Sales Call Summary (Prospect Type, Products Discussed, Number Non-Complainces) |
“As a QA Analyst, I need unified guidelines for each product for fair evaluations” | Fragmented compliance guidelines for multiple products cause confusion | Smart Contextual Guidelines |
“As a QA Analyst, I want to prioritize tracking compliance to boost productivity” | Multitasking during call evaluation reduces analysts' productivity | Optimized Evaluation Workflow |
Solution
First of it's kind AI-Powered QA Platform for evaluating Enterprise Sales Calls
A fragmented, tedious process that took 60 mins to complete was shrunk to 5 mins using the integrated QA platform
Optimized End to End Evaluation Flow
The final interface delivers three core pillars: an allocated call list view with advanced filters, upfront key data points to eliminate ambiguity, and a streamlined review screen where analysts review NLP-flagged non-compliant practices rather than manual audio playback
Designs
The Solution in Action: Designed for Speed and Accuracy
A list view of allocated calls, coupled with relevant filters, helps QA Analysts find specific calls in seconds
Presenting key data points upfront eliminates ambiguity about the nature of the call
QA Analysts can skip identifying non-compliant practices and focus on reviewing non-compliance flagged by the platform
Impact
Transforming an Operational Bottleneck into a Scalable Enterprise Engine
Business Impact: Successfully scaled weekly call volume by 11x, bringing the enterprise into full regulatory compliance without increasing operational overhead.
Consulting Recommendation: The platform's AI & NLP capabilities can transform it into a Sales Strategy and Training tool for the organization
Sneak Peak into other users of the platform
Call Evaluation isn't the entire story. A QA Manager is responsible for allocating calls and managing the teams performance. Here's a sneak peak into their task
Want to know the messy part? (How we drilled down on the core problem? What were some non-negotiables? Concept pivots due to technical constraints) Let's chat!
hiremath.rishikesh@gmail.com
