



Client
Top U.S. Legal Tech Firm
Challenge
Auditing phone calls and SMS messages was manual, slow, and inconsistent, taking days or weeks.
Solution
An auditing AI agent system transcribes calls, classifies communications, and batch-processes thousands of records.
Overview
A leading legal tech company partnered with StackAI to automate auditing of external communications. In just weeks, they replaced manual transcription and review with a secure, AI-powered workflow that reduced QA costs by 64%, cut operational workload by 92%, and scaled oversight to 100,000+ calls per month—all without adding staff.
64% reduction in QA costs
92% workload reduction for the operations team
100,000+ calls and messages audited monthly without additional headcount
50% fewer errors compared to manual review
The Problem: Manual Audits Created Risk and Delays
Despite having a robust CRM, the company’s operations team was overwhelmed by the manual auditing of calls and text messages. Each transcript required rule-checking and quality scoring—a process that was slow, inconsistent, and error-prone.
Audits took weeks to complete and covered only a fraction of communications, creating gaps in oversight and exposing the company to compliance risks. Leadership needed a faster, more scalable approach that did not compromise data privacy or accuracy.
50% Fewer Errors. Faster, Smarter Audits at Scale.
StackAI reduced the firm’s audit turnaround time from days to hours. What once took 14 days now takes just 14 hours for voice, and SMS auditing dropped from 3 days to 3 hours. Overall time spent on QA fell by 92%, while error rates dropped by 50%.
Thanks to AI agents, the company has scaled its audit scope dramatically, analyzing 100,000+ calls per month without hiring additional staff. The financial upside was just as impressive: with over $60K in monthly cost savings, the company improved oversight while cutting operational expenses.
The Solution: Secure, AI-Powered Auditing With StackAI
Using StackAI’s no-code platform, the company built a secure, multi-step AI auditing assistant:
Accurate Transcription: Phone calls were transcribed with Deepgram’s speech-to-text models.
Automated Evaluation: An Anthropic LLM in AWS Bedrock analyzed transcripts and SMS messages against internal compliance and quality rules.
Batch Processing: A user-friendly interface allowed teams to upload and process hundreds of communications at once.
This end-to-end agent reduced turnaround time from 14 days to 14 hours for calls and from 3 days to 3 hours for SMS, delivering faster and more reliable audits at scale.
“AI agents transformed every step of the team's quality control process, making it simple to implement and scale AI systems. They can now process 10 times more calls and messages with a higher degree of accuracy, and much more cost-effectively.”
Yiyi Wang
Forward Deployed Engineer
Faster, Smarter Oversight
The AI auditing system transformed quality control:
• QA review time dropped by 92%
• Error rates decreased by 50%
• Monthly cost savings exceeded $60,000
• Audits expanded from a limited sample to 100,000+ communications monthly
By automating this critical but resource-heavy process, the company achieved stronger compliance, lower costs, and dramatically higher audit coverage, all while keeping sensitive data secure. The success of this workflow is now inspiring the company to explore new AI use cases across compliance and customer-facing operations.
Customers
Explore More Customer Stories
From Weeks of Research to Minutes: How NobleReach Became the AI-First Nonprofit Leading Tech Transfer Innovation
Manual research, competitor analysis, and tech transfer reports took a full week per project, slowing impact and overwhelming a small nonprofit team.
How Varos Saved 800+ Hours With an AI-Powered Categorization Agent
The operations team spent countless hours manually scanning company profiles, analyzing product offerings, and categorizing leads
How Nova Talent Cut Recruiting Costs by 60% With AI-Powered Application Review
Nova Talent’s admissions process required recruiters to manually review resumes, written answers, and video interviews