Customer story · Professional Services
SupportZebra sees 78% drop in customer noise complaints
SupportZebra, a Houston-based outsourced customer support provider, equipped its work-from-home agents with Krisp noise cancellation during the COVID-19 pandemic and cut customer noise complaints 78% with no hardware changes.

About SupportZebra
SupportZebra is an outsourced customer service and technical support provider headquartered in Houston, Texas. Founded in 2010, it helps SaaS and ecommerce businesses scale by providing customer service and tech support through talented, highly English-proficient call center agents. The company has more than 500 team members, and during the COVID-19 pandemic it moved 90% of its staff to a work-from-home model.

The challenge
SupportZebra's business relies on phone calls. As a provider of outsourced customer support for SaaS and ecommerce brands, it needs quality audio for its call center agents to do their jobs well.
When the COVID-19 pandemic forced SupportZebra to move 90% of its staff to a work-from-home model, the fast-growing company faced a thorny problem: more background noise and, with it, more customer complaints. Home can be an unpredictable and noisy place. Barking dogs, talkative children, even a fan blowing could spell disaster for a customer call taken from an agent's home, and the rise in complaints threatened the viability of the business. As Founder and CEO Nathan Yap put it, the number of customer complaints skyrocketed due to background noise being heard.
SupportZebra needed a way to keep its workforce safe at home while still keeping customers satisfied, with a solution that was easy to implement regardless of an agent's technical skill and that worked with the headsets, microphones and communication apps already in use.
Agents, their managers, and customers are all ecstatic and happy, and the difference is black and white.
The solution
SupportZebra equipped its agents with Krisp noise cancellation, which integrated with the call center agents' existing equipment to cancel background noise and deliver clear audio on calls. Because Krisp's technology is bidirectional, it mutes background noise coming from both the agent's end and the customer's end, so both sides get a high-quality call experience.
Krisp's AI-powered technology has been trained on more than 20,000 sounds and 50,000 distinct speakers to pinpoint and isolate background noise so that only the intended speakers' voices come through. Processing happens locally on the agent's device, so no audio is leaked and privacy is protected. Unlike noise cancellation apps that distort voices, Krisp preserves the integrity of the speaker's voice, avoiding robotic-sounding audio.
Krisp is a software-only solution that supports all call center equipment and CRMs, works with any headset, microphone and speaker, and integrates with more than 800 communication apps. With more than 500 team members, most working from home, SupportZebra found the transition seamless: the app supported all of its headsets, microphones and speakers and integrated with its communication apps, with no special technical skill required.
As the company returns to the office or moves to a hybrid model, SupportZebra plans to keep using Krisp to maintain a noise-free calling environment whether agents work in the office or at home.

The results
- 78% decrease in customer noise complaints
- Agents kept working safely from home while delivering the same quality of support as in the office
- Two-way noise cancellation removed background noise from both the agent's and the customer's side of the call
- Seamless rollout across existing headsets, microphones, speakers and communication apps, with no technical skill required
- The document's results panel also lists 25% higher first call resolution, 16% more sales conversions, 10% lower AHT, 20% lower call abandonment
- The same panel lists 8% higher CSAT scores, 25% higher agent CSAT and 30% lower cost

Get the full case study
Download the original document as published by Krisp (PDF, 1.1 MB).
Your story could be next
Want results like these?
Every result on this page started with one conversation — let's have yours.


