Level AI’s Impact on Contact Center Call Categorization

Contact Center Call Categorization
Share this article
Share on facebook
Share on twitter
Share on linkedin

When I was part of the Amazon devices (Kindle, Fire tablets, etc.) team, our contact center team was the hub of customer engagement and troubleshooting. 

To scale a budding business, we wanted our customer service call center agents to disposition call reasons accurately for analytics and deep dives.

We faced several call center categorization and agent training challenges. Amazon used a common approach to call categorization. 

The agent selected the call category from a tiered tree at the end of every contact. This approach had two primary issues.

  1. Definition and management of call center categories
  2. Accuracy of contact categorization

Improving Call Categorization Definition and Management

The first issue was the definition and management of the categories. We went through a few iterations of expanding the number of categories to gain issue fidelity. 

Later, we reduced the number to simplify. Every change required new guidance and training for agents, and it never resulted in feeling more confident in the data.

Improving Call Categorization Accuracy

A study by our quality improvement team found that many contacts weren’t being categorized well. 

Poor categorization was not surprising because agents weren’t rated on the quality of their categorization. An experienced team lead performed this review.

Instead, they were rated by their average handle time (AHT), and searching for the perfect category takes time. 

Additionally, the study also found that most contacts had multiple categories. Unfortunately, we forced a single category for the call and lost a lot of rich data in the process.

Benefits of Improved Contact Center Call Categorization

These are the same problems we see teams of all sizes struggle with in the modern contact center. 

What we needed was an analytics layer that could categorize calls appropriately. An analytics layer would have: 

  • Helped reduce after-call work leading to more calls handled
  • Improved data quality and call center metrics
  • Kept agents focused on helping customers and improved the customer experience 

Even better if the system could help identify categories and provide real-time assistance to agents in finding the correct information for customers. Call center performance and agent performance would increase, and the service level could also be improved.   

I am excited about Level AI because it was clear how the product addresses these challenges. Level AI gives superpowers to your support organization. 

At the product’s core, the semantic intelligence engine helps identify detailed drivers of call interactions, not just the words, with your customers. It enables your agents, quality assurance team, and product teams to operate more efficiently and with greater insight.

What to do Next: Contact Center Call Categorization with Level AI

Let’s talk about how we can help your business get to the next level. 

Here’s what you can do:

Get Your Personalized Demo

Augment your agent and QA team performance with an intelligence system for the modern contact center.
Please Fill in the Form to Request your Demo

How To Evaluate AI for Modern Contact Centers