1% Visibility to 100% Understanding: How HTC Transformed QA with Level AI

HTC’s vision for member experience

HTC isn’t just another telecom provider; they are the nation’s largest cooperative, distinct in their mission to serve the community rather than just shareholders. Their reputation hinges entirely on the quality of the member experience provided by their 100 member-facing agents, who manage over 200,000 complex interactions annually across Technical Support, Residential, Retail, and Business services.

But for Lana Suggs and her team, maintaining this high standard was becoming a mathematical impossibility. They wanted to coach effectively, but their visibility was blocked by manual limitations.

"We are the nation's largest cooperative... and our member experience is exceptional. But we wanted a way to be able to measure that more accurately... so [specialists] could actually speak to and coach more efficiently."

— Lana Suggs, HTC

The "Good Job" trap: The limits of manual sampling

For years, HTC’s Quality Assurance (QA) process relied on manual spreadsheets. The QA team and supervisors were dedicated, but the process only allowed a small fraction of agents’ interactions—1% of total calls—to be reviewed for annual evaluations.

On paper, results looked impressive: most agents scored in HTC’s top tiers, the Blue or Purple range. But Lana knew these scores didn’t tell the full story. “A small sample doesn’t capture the day-to-day reality of performance,” she explains.

Finding calls to review was time-consuming for both QA evaluators and supervisors, and the limited sample sometimes left agents feeling only their toughest calls were being selected.

HTC transitioned to a more comprehensive QA system, reviewing a broader set of calls. Now, feedback reflects overall performance, and both evaluators and supervisors can focus on coaching and development rather than just tracking down recordings.

The shift transformed QA from a “mirage of perfection” into a meaningful, actionable process that supports agents, empowers evaluators and supervisors, and ensures every customer interaction is evaluated fairly.

"There was a perception of, 'My toughest call is going to be selected.’ That’s why
it was important for us to share with our agents that their manual evaluations are
now randomized... our QA specialists and supervisors are not cherry-picking
these calls."
— Lana Suggs, HTC

Replacing Guesswork with Analytics

HTC needed a partner that could integrate with their specific Mitel platform and provide the analytics necessary to see the full picture. They selected Level AI not just for automation, but for its ability to turn raw data into coaching strategy.

"A major factor in selecting Level was its forward-thinking and progressive approach... Your platform was very robust and the integrations made that very easy for our team." — Lana Suggs, HTC

1. Solving the Visibility Gap with Insta Score: To shatter the "1% blind spot," HTC
   switched from manual sampling to 100% automated scoring. Level AI’s Insta Score
   began grading every single conversation instantly. This moved the team from a lottery system to a comprehensive review of daily performance, ensuring that "one tough call" didn't define an agent's entire month.
   
   

2. Solving the Efficiency Gap with Deep Analytics: The shift to Level AI exposed deep operational insights that spreadsheets never could. Phoebe, HTC’s subject matter expert, utilized the Level AI Analytics and "Workers" tab to revolutionize her workflow. Instead of spending hours listening to random audio files, the system now auto-generated trends and coaching plans for her. "It uncovers trends automatically versus me trying to find them manually," she notes. This analytics-first approach allowed her to spot systemic issues instantly.
   
   

3. Solving the Trust Gap with Auto-Assignment: As HTC modernized QA, transparency became a top priority. Previously, evaluators and supervisors manually selected calls, which sometimes left agents unsure if reviews reflected their day-to-day work.
   
   With Level AI’s auto-assignment, calls are now randomly selected based on objective criteria, ensuring a fair and unbiased sample. Agents can trust that every review reflects their true performance, while evaluators and supervisors focus on meaningful coaching.
   
   

4. Solving the Vendor Mystery with VOC Search: The power of the platform extended beyond coaching. When agents began reporting vague error messages from a third- party vendor, the team used Level AI’s Voice of Customer (VOC) keyword search to hunt for the phrase "something went wrong." The system instantly quantified the exact volume of the error, giving HTC the hard data they needed to force a resolution.

The Value of "Green": Achieving True Performance Clarity

The transition to Level AI didn’t just change the workflow—it aligned HTC’s metrics with reality. The impact was measurable and immediate:

• Doubling Manual Review Capacity: Beyond automated scoring, efficiency gains allowed the team to double the number of deep-dive manual reviews, greatly expanding hands-on coaching opportunities.
  
  

• From 1% to Full Coverage: Moving from reviewing 1% of interactions to scoring
  100% of calls via Insta Score removed guesswork from performance management.
  
  

• Realistic Performance Recalibration: Artificially high “Blue/Purple” scores, previously held by the majority of agents, shifted to a realistic median “Green,” giving agents a genuine baseline for improvement.
  
  

• Coaching Prep Reduced to Seconds: Time spent hunting for call recordings was eliminated, enabling supervisors and QA evaluators to generate coaching plans instantly using the “Workers” tab.

“Before, we were reviewing 1% of calls. Now, with full scoring and doubled manual reviews, most agents are in ‘Green,’ giving us a real picture of their day-to-day performance.” – Phoebe, HTC