Reducing Call Center Operational Costs 

Overview of the Project 

In a rapidly evolving insurance landscape, we partnered with a leading insurance provider to tackle a critical business challenge: high operational costs associated with their customer call center. The project focused on leveraging data from their enterprise data warehouse (EDW) to better understand why their members were calling, accurately predict call volumes, forecast costs of onboarding new groups, and enhance the overall member experience.  

Business Problem 

The Finance & Operations teams were looking to reduce the costs of their call center operations.  

Their traditional reporting operations were burdened by not fully understanding why members were calling, which members are likely to call in the next 30-days, what events occurred before and after a call, how much it costs to onboard new groups of customers, and, ultimately, inaccurate forecasts for future call volumes.  

As a result, the client receives tens of millions of calls per year which equates to over hundreds of millions in operational costs.  

Solution Approach 

As experts in guiding clients through the insights journey, the expertise of the CloudEQS team enabled the client to go from reporting to understanding to anticipation. Here was our multi-faceted approach to tackle the client’s challenges: 

1. Data Integration: The CloudEQS team worked with leading data integration provider, Matillion, to ingest raw data from structured, semi-structured, and unstructured data across members, demographics, enrollments, claims, grievances, and call centers data sources. 

2. Central Source of Truth: With Snowflake’s Data & AI Cloud platform as the destination, we developed a centralized data platform that easily integrated these disparate datasets into a central repository.  

3. Customer Segmentation & Clustering: In order to progress the client to the advanced analytic & data science phase, we created 10 clusters of customers with similar attributes & characteristics. Group features that made up the clustering included total calls, claims, website visits, denied amounts, contract type, customer category (e.g. high touch, low touch), claims by age, and state-wide customer counts.  

4. Predictive Modeling: By applying predictive data science models against the historical data, we were able identify reasons for calls, which cluster of customers were most likely to call, identify patterns & trends, and predict forecast future call volumes for the next 30-90 days. 

Results 

With the new level of insights, the business outcomes include:  

• Accurate Call Predictions: The client can predict expected call volumes in the next 30,60, & 90 days with 99% accuracy.   

• Staffing Cost Reduction: With this new level of insights, the client is able to accurately staff their call center and reduce the agent’s idle time.   

• Accelerate Training & Knowledgebase: With the new insights about their members, the client is able to have a more up-to-date knowledge base. Resulting in improving the training for call center employees. 

• Enhanced Pricing Strategies: With deeper insights into their members’ experience and customer profiles, the client can refine their pricing strategies for their service offerings for similar customers.  

Conclusion 

This project exemplified the power of a robust & modern data platform, data engineering, and AI/ML in transforming operational efficiencies within the insurance sector.  This underlines the importance of accurate and timely data, which forms the backbone of the AI/ML initiatives and enables the precise predictions and efficient operations described. By addressing the client’s call center challenges, we delivered substantial cost savings, accurately predict future outcomes, and enhanced the overall member experience.