Problem

Call centers are a crucial part of customer service, but it can be difficult to assess the quality of calls and ensure that agents are providing excellent service. Traditional methods of call assessment involve listening to hours of audio recordings, which is time-consuming and inefficient. Additionally, it can be challenging to identify trends and patterns in call data.

Our team was tasked with finding a solution that would streamline the call assessment process and provide valuable insights into agent performance.

Watch this video

Solution

We created an AI-powered platform that assessed calls using NLP techniques. The platform split each call into six sections across two channels (agent and customer). We then used topic modeling to assign a topic to each section, allowing us to quickly understand what was being discussed without having to listen to the entire call.

In addition, we used sentiment analysis to evaluate the customer journey and determine if agents were providing excellent service. We grouped calls based on various filters (client, region, country) and compared agent performance within teams. We also assessed agents using transcript calculations such as Overtalk (speaking over the customer) or increased silence (awkward pauses), ensuring they maintained an appropriate pace of speech.

Impact

Increased insight capability by allowing quantification of emerging themes and topics on calls
Generated even more revenue offering insights and consulting to clients who are outsourcing their call centres.
Reduced time spent on call assessments by 90%, while assessment coverage of calls skyrocketed from 1% to 100% of inbound calls.
Reduced Full Time Effort of agents and focused on calls
Improved Focus Training of lacking agents on specific areas

"This platform has revolutionized our call center operations and allowed us to provide better service to our customers. We are thrilled with the results."

Team

George Petropoulos - Data Scientist
- Developed NLP techniques for call assessment
- Implemented sentiment analysis for customer journey evaluation
- Topic Modeling for theme detection over calls
Chris O. - Project Manager
Titus M. - .NET Dev Manager
- His team was able to build the Front-end to bring up all the implementations to life to the end user.

AI Streamlines Call Center Assessment & Boosts Performance