Reduced Unproductive Truck Rolls for a Telecom Provider, cutting costs and improving response times to boost customer retention

About the client

A US Based Telecom Provider

Challenges

• Approximately 25% of truck rolls (dispatches) were deemed unproductive, significantly increasing the cost per service.
• Unproductive dispatches led to longer and unreliable response times, resulting in customer dissatisfaction and loss of customer base.

Approach

• Data Integration and Analysis: Collected and harmonized data from 25+ disparate sources covering over 100 variables, including technician logs and trouble ticket information.
• NLP-Powered Insights: Leveraged Natural Language Processing (NLP) to structure unstructured data such as technicians' remarks and trouble ticket details for deeper causal analysis.
• Predictive Analytics: Built machine learning models to identify key drivers of unproductive dispatches and recommend proactive mitigation strategies.
• Real-Time Alerts: Designed and implemented an alert system to notify teams about potential issues before dispatch.

Solution

• Integrated data from multiple sources and identified key causal factors contributing to unproductive truck rolls.
• Triggered proactive alerts based on insights from predictive analytics to enable timely mitigating actions.
• Streamlined data processing workflows using NLP-powered automation, improving the structuring of unstructured data.