Company Overview

Resident is a direct-to-consumer leader in the mattress and home goods space, known for its 365-night trial and customer-first approach. With over 550 employees and $180M in annual revenue, Resident was recently acquired by Ashley Home in a $1B deal. As the company scaled, so did the complexity of delivering consistent, retention-driving experiences across every customer journey.

The Challenge

CSAT and manual dispositions tagged by agents painted a clear picture—but it turned out to be the wrong one.

CSAT surveys and agent-applied tags flagged “returns” as the top driver of dissatisfaction. On paper, the conclusion was obvious: the return process needed fixing.

But something didn’t add up. Resident’s analytics team dug deeper and found that most customers were actually satisfied with their return experience. Still, these interactions had disproportionately high negative sentiment, escalation rates, and DSAT scores. So what was going on?

The team realized they were chasing ghosts in the data:

• Agent-applied tags were inconsistent and error-prone, skewing categorization and masking root causes.
• CSAT response rates hovered around 20%, offering only a partial view of customer sentiment.
• The company lacked a reliable way to connect customer emotion, policy friction, and downstream business impact.

What looked like a process-wide issue was actually a localized friction point—one step within the return flow that was rigid, hard to navigate, and causing unnecessary customer frustration.

The real risk? If Resident had updated its return policies based on these flawed signals, it could have made things worse—escalating costs, reducing LTV, and missing high-value save opportunities.

What the team needed wasn’t more surveys or tags. They needed a scalable, trustworthy way to automatically classify customer conversations, surface hidden patterns, and tie those insights to business outcomes.

The Turning Point

Resident turned to MaestroQA to uncover the story behind the numbers.

They began by building AI-powered LLM prompts to classify every customer conversation by intent—such as returns, cancellations, or delivery issues—so they could move beyond anecdotal assumptions. When sentiment, CSAT, and handle time were layered in, a clearer picture emerged.

Returns, it turned out, weren’t the core issue. A specific step in the return process was disproportionately driving dissatisfaction. Once the team isolated that moment with data to back it up, they updated policy and retrained agents—leading to fewer escalations, improved CSAT, and better retention outcomes.

“We actually learned that returns are extremely successful... there was just one part of the return process that was causing the friction.”

The Solution

A Scalable Framework for Voice of Customer

Resident built a three-part LLM-based tagging system that categorized conversations by:

• Intent: Why the customer contacted support
• Solution Offered: What action the agent took
• Outcome: What happened next (retention, churn, escalation, etc.)

They created 24 custom “dispositions” to reflect key customer journey moments, replacing inconsistent manual tagging with more accurate and consistent AI labeling.

By integrating this data with operational metrics—such as handle time, DSAT rate, and repeat contact volume—Resident was able to uncover friction across the entire customer journey and take action across departments.

Impact

Returns → CSAT Rebound & Retention Gains

Resident used conversation data to pinpoint a policy-level issue within the return process. Once identified, they updated the workflow and retrained agents. CSAT scores improved, and fewer customers churned after initiating a return.

Transit → Faster Recovery, Fewer Cancellations

Transit delays weren’t the most frequent issue, but they had the highest negative sentiment. Using MaestroQA, Resident identified the top frustration point and worked with their transit partner to offer faster recovery. This led to fewer canceled orders, escalations, and repeated contacts.

WFM → Handle Time Cut in Half

Conversation tagging revealed that agents authorized to resolve issues without lead approval had 50% lower average handle time. Based on that insight, Resident expanded agent certifications—improving efficiency and customer satisfaction.

RCA for Any Team

When contact volume spikes or forecast mismatches arise, teams across Resident—from workforce to logistics—use MaestroQA to diagnose the root cause. They no longer rely on guesswork, they use conversation data to make decisions grounded in volume, sentiment, and intent.

Resident’s transformation didn’t come from grading more tickets. It came from using AI to listen better—and act faster. By turning conversation data into operational intelligence, they improved CSAT, increased retention, and made their policies more profitable.

What’s Next

Resident is now expanding the framework to track not only intent, but also solution types and conversation outcomes. This will help tie support interactions more directly to retention and lifetime value.

They are also exploring a Customer Experience Score to complement (and eventually surpass) CSAT, measuring every interaction with more consistency and nuance.