NOTE: Demo visuals use either blurred real data or synthetic placeholders to protect customer privacy.
Manual Audits Limit Quality and Scale
Exceptional customer support is one of Abnormal’s greatest strengths. But ensuring consistent quality across every case takes time, especially when each audit requires manually reviewing ticket history, documentation, and handling standards.
To streamline this process, Thane Allgood developed the Customer Support Quality Assurance (QA) Glean Agent: an AI-powered auditor that transforms how Abnormal evaluates support interactions.
Every week, Abnormal’s support team handles dozens of customer cases, each requiring careful resolution and follow-up. To maintain excellence, leadership mandated thorough quality inspections across all support interactions.
Manual QA was slow and labor-intensive:
- Each review took 20–30 minutes, as auditors manually compared tickets against internal standards.
- Reviewing enough cases to get a meaningful sample size required dozens of hours per week.
- Valuable time spent on auditing meant less time focused on enablement and coaching.
Thane identified that we needed a way to inspect support quality at scale without burning hours on manual reviews.
Automated Case Audits with Glean
The Customer Support QA Agent, built in Glean, automates nearly the entire audit workflow.
Here’s how it works:
- Input a case: The auditor opens the Glean agent and enters the case URL and the name of the Technical Support Engineer (TSE) being reviewed.
- Data gathering: The agent automatically pulls all relevant case data, including correspondence, documentation, and associated performance metrics.
- Apply QA rubric: It references Abnormal’s internal case-handling best practices and performance requirements.
- Analyze and score: The agent evaluates the case against the rubric, identifies strengths and weaknesses, and calculates a quality score out of five.
- Generate feedback: The output includes a detailed breakdown of what the TSE did well, what could be improved, and recommended next steps.
In Thane’s demo, he audited a case handled by one of the team’s top performers. The agent returned a perfect 5/5 quality score, citing adherence to best practices, clear communication, and timely resolution, all within minutes.
The process, which previously took half an hour, now takes under five minutes and produces a structured, repeatable report every time.
From Manual Review to Continuous Insight
The QA Agent not only speeds up case review, but also standardizes quality measurement across the organization.
- Consistent scoring: Every case is measured against the same rubric, eliminating reviewer bias.
- Scalable auditing: The team can now perform three times as many audits each quarter without adding headcount.
- Actionable feedback: Each audit generates insights that can be fed back into Glean for trend analysis and coaching.
- Foundation for automation: The current agent is the first step toward a fully automated QA system that can audit every case daily.
This takes us from burning man-hours on inspection to spending those hours on enablement, helping TSEs grow instead of checking their work.
Higher Throughput, Faster Feedback, Better Support
The Customer Support QA Agent has already had a measurable impact:
- Audit volume tripled in one quarter.
- Review time per case was cut by over 80%.
- Consistent performance tracking enables targeted training and recognition.
- Faster insights mean support leaders can spot systemic issues early and adjust processes proactively.
With these gains, the support organization is moving toward a future where quality oversight happens continuously and automatically.
Fully Automated QA at Scale
The next phase of Thane’s project focuses on continuous improvement and automation. The team is already:
- Stress-testing the agent to measure accuracy and variability in scoring.
- Calibrating performance using telemetry data to fine-tune consistency.
- Breaking down workflows to improve speed and precision with each iteration.
- Exploring full automation, where the agent runs on a cron job to audit all closed cases daily.
More automation means higher throughput, broader visibility, and faster learning loops, creating a self-improving feedback system for the entire support org.
What Makes The Customer Support QA Agent Awesome
Thane’s agent is a prime example of Abnormal’s innovation culture in action. By automating a repetitive, time-consuming process, he built a system that not only improves efficiency but also enhances quality and consistency.
It’s a simple idea with profound impact: let AI handle the audits, so humans can focus on improving the customer experience.
