Purple Innovation Uncovers Critical Customer Insights in Hours Instead of Weeks

Science-Based Comfort Meets Data-Driven Care

Purple isn’t just a mattress company; they are engineers of comfort, famous for their unique GelFlex Grid technology that changed how people sleep. Based in Utah, their mission is to help people feel better and live better through science-based sleep products.

Keeping thousands of customers happy requires listening closely to what they say. Leading this charge are Jane Greiner, Director of Operations of Customer Care, and Angie McDonald, Director of Analytics and Workforce Planning. Together, they ensure that every customer interaction feeds back into the business to improve products and services.

"We're a science-based comfort company focusing on mattresses and pillows, and other wonderful sleep products." — Jane Griener, Director of Operations of Customer Care

Drowning in Spreadsheets

Before Level AI, getting answers from customer data was a struggle. The team was relying on a legacy speech and text analytics tool that felt like a "black box." It required such specialized technical knowledge to navigate that it often felt like they needed a separate degree just to run a simple report.

The real pain of this system became obvious during a major product launch. Looking for insight into the sales of the new product, leadership asked for a "deep dive" to understand customer behavior.

In a modern world, this should have been a quick query. Instead, it triggered a manual marathon. The team had to download raw call results and assign a team member to look through each interaction one by one. They spent days manually marking up spreadsheets, trying to spot trends in a sea of noise. While they eventually found the answers, the process took weeks, time the business couldn't afford to lose.

They had the customer conversations, but without a way to unlock them quickly, potential revenue could be slipping away.

"We felt like we needed to have someone who almost had a degree in how to use the tools that we had... getting insights out of those interactions was really manual and took us lots of hours, days, you know, sometimes weeks." — Angie McDonald & Jane Griener

The Mystery of the "Too Tall" Mattress

Purple needed a change, and Jane & Angie took charge to bring this change. They evaluated several vendors, but didn't want another tool that required an expert to run; they wanted answers. Level AI stood out for its mix of simplicity and power. It was incredibly user-friendly yet empowered the customer care and analytics teams to dive deep into data and pinpoint root causes without heavy manual lifting.

The implementation was smooth and fast. "Honestly, I was really happy with our implementation. I thought it was one of the better managed implementations that I've been a part of," Jane noted.

Building a Smarter Foundation to maximize impact, the team didn't just turn the software on; they strategically configured it to capture high-value data.

• Targeted Tagging: They created specific tags to capture important business scenarios and product SKUs, setting up their database for powerful, granular reporting.
• Fairer QA: They revisited AutoQA rubrics to refine questions that felt penalizing to agents, ensuring the system was fair and constructive.
• Holistic Listening: They integrated Voice of Customer (VOC) data from surveys and product reviews to gain a comprehensive understanding of customer sentiment.

Once Level AI was live, the team started solving mysteries that had previously baffled them.

One standout moment involved a new high-end mattress - a direct result of their new SKU-wise VOC insights. Customers kept raising the bed’s height as an issue, which surprised Purple’s product team—at 17.5 inches, the mattress was tall, but should not have been an issue for their customers.

Angie's team used Level AI to analyze the conversations, and it turned out that delivery partners were setting the adjustable bases to the highest possible setting during installation. Combined with the tall mattress, the beds were ending up too high for the customer to comfortably get into.

Because the team quickly identified this trend with the Level AI Conversation Intelligence suite, they pinpointed a specific training gap. The team realized that they needed to communicate better with customers and delivery teams to ensure the bases were set to the lowest profile.

Purple’s team also used the tool to increase customer confidence during the Black Friday promotion period. Level AI identified a trend where customers were hesitant to buy, fearing a better deal would drop on Black Friday. The team immediately spun up a plan to educate and reassure customers that Purple always honors any future price drops that occur within 30 days of their purchase, giving buyers the confidence to purchase right then and there.

"We even made a couple of suggestions when we were trying things out, and those suggestions have already been implemented... With Level, it's coming out the next quarter, so I love that." — Jane Greiner

100% Coverage and Faster Answers

The difference in speed and efficiency has been night and day. Tasks that used to take a small army of people weeks to complete are now finished in a single afternoon.

The shift to Level AI delivered quantifiable improvements in Quality Assurance and agent performance:

• 100% QA Coverage: Purple scaled from partial checks to 100% QA coverage across all voice channel conversations over 2 minutes.
• Evidence-Based Training: The wealth of data allows them to provide specific, evidence-based feedback to agents, directly increasing First Contact Resolution (FCR).
• Instant Product Feedback: Search Worker insights now provide immediate feedback on new product launches, enabling the product team to respond in real time.

This efficiency came at the perfect time for Jane. Purple has committed to efficiency gains as a high-priority initiative. Despite having a leaner department, she hasn’t lost visibility. With Level AI, she hasn't missed a beat. She can dig just as deep, if not deeper, without needing extra hands.

The frustration of waiting "maybe until 2026" for vendor features is gone. They now have a partner that moves as fast as they do.

"In the past... we had to download all of the results and go through call by call... But in Level, I'm doing something similar now and can do it in a couple of hours. As long as my structure is set up correctly and my tags are all great, I can do it in just a few hours. I have fewer resources than I had in the past, and I can still deliver the insights with the Level AI that I could before." — Angie McDonald