man sitting in an office working on his computer

Triage Customer Support Calls

Create accurate estimates of call time and likelihood of success based on customer demographics and type of call.

Time to resolution for customer support calls is an important metric in optimizing call center performance. Through predictive modeling, customer service teams can estimate how long individual calls will take and assign them according to the availability of representatives.
image

Efficiency Gains

Automatically route calls to appropriate reps based on difficulty, type of call, geographic and demographic data, and employee availability
Automation_icon

Bottom-Line Returns

Continually report and optimize call center performance automatically
Image

Customer Experience

Affect time to resolution and customer sentiment and fully address customer needs

Business Problem

The Time to Resolution (TTR) metric is front and center for most customer support teams. It measures the amount of time that elapses from the moment a support case arrives in the customer support queue until the case has been resolved. At intake, an administrator assesses the case, measures the workload of support representatives, and assigns the case to the representative best qualified to solve the case in the shortest time. It no longer suffices to manage support cases with spreadsheets and anecdotal data. Nor is it best practice to let a phone queue or email inbox treat all cases on a first-come, first-served basis. Growing companies face customers with rising expectations for prompt, authoritative, speedy support. Traditional ways of performing triage on customer support inquiries are too slow and unreliable, putting the customer experience in jeopardy.



Analytics Solution

A company can accelerate and improve triage by using a predictive model to reduce TTR. The model is an application of augmented machine learning that bases predictions about TTR on historical case data. The predictions take the availability of each support representative into account and determine which one will be most likely to resolve the issue quickly. The more data the company has available for training the model, the more accurate the resulting predictions. Sample workflows include steps for initially preparing data, generating and appending additional variables, removing outliers, imputing data, and testing. Through machine learning, the model shifts the labor-intensive component of vetting cases off of humans and onto computing resources. It gets inquiries into the hands of customer support representatives promptly, reducing overall TTR and ultimately providing quicker and higher-quality service.

With Alteryx, you can:

  • Use Input Data tool to load in historical customer support data or use a Connector to automatically pull in data
  • Automate joining and blending of data, including engineering new features to load into Alteryx Intelligence Suite tools for advanced analytics
  • Connect created models to customer support platforms, or feed into Tableau dashboards for historical and predictive customer support visualizations
Customer Support Workflow

1 - Data Connection

Data Connection: Combine historical customer support data and customer demographic data

2 -Prep & Blend

Combine data from various sources and prepare data for analytics

3 - Advanced Analytics

Feed cleansed data into advanced analytic workflow modules or customer support platforms

Additional Resources

データアナリスト

顧客分析スターターキット

電子ダッシュボード

Tableau スターターキット

people working in an office answering calls

Reduce Support Case Time to Resolution

woman talking on a headset

Call Center Analytics Predictive Modeling

colleagues white-boarding

Customer Journey Analytics

分析を変革

データに潜むインサイトを引き出しませんか?

無料トライアル