RapidMiner, Inc. All rights reserved. - 1 - TM Call Center Chat Analytics Achieving actionable insights from customer interactions in a Service Center Surya Putchala Lead, Big Data Analytics Cappius Technologies Tom Ott Marketing Data Scientist RapidMiner
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
TMAgenda• Introduction to RapidMiner• Business Drivers for Call Centers• Call Center Chat Analytics– Chat Process & Solution Architecture– Algorithms and Methodologies
The key to innovation and competitive advantage lies in the power of data science. Invest in the right tools and the right skills to uncover new opportunities – and change the
The dynamics of the call center are changing because of the adoption of chats and other text channels. The drivers are :• Smartphone usage for chat• Multi-task chat windows• Reducing Service agent visits • Cheaper cost
Real-time Speech Post ChatPost Call Speech Live Chat Support Email Social Media
The chat is set up for each customer. It captures name of the user, product/service the customer wants and, specific issue about the product.
Pre-chat Info Chat Post chat Survey
The post chat survey are direct questions about the chat experience, if they would want the chat forwarded, reasons for dissatisfaction and overall satisfaction.
Information from Chat Meta-data Total lines (Operator/Customer ) Agent-Customer Response Times, Handle times Survey statistics Characteristics (Volumes, averages, abandoned ,
duration, word counts /lines by customer/agent) Utilization (engaged time, idle time, online time)
We consider the last three sentences of the customer to be very important predictor of his satisfaction and has a great predictive power of deciphering customer satisfaction. We weigh this factor higher.
There can be multiple themes in a chat conversation. The mood, sentiment, tempo of the conversations could be modelled as a network. This is inferable factor but has a great impact on arriving at a sentiment score.
AveragesWeighted AveragesSmoothing
The ability of an agent to convince the customer and close an issue successfully with a customer can yield a great customer satisfaction. Hence the sentiment of a Chat is also a predictor for Customer Satisfaction.
a. In memory processingb. Large Scale Data processing (Clustered Environment)
2. Maintainability and Model Management3. Developer Productivity and Environment4. Integration and Data Persistence5. Capabilities
a. Connectors to variety of Data Sources (Input / Output formats)b. Ability to extensively Cleanse and Munge datac. An extensive array of Modeling Techniquesd. Built in Schedulers ( For data ingestion and report generation)e. Extensibility of functionality
TMKey Takeaways Chat Analytics solution addresses call center challenges
through Sentiment Analysis, Agent Scoring and Topic Detection
Improves productivity by rapidly surfacing customer and business issues and opportunities
Serves as an early warning system to identify issues before they escalate and impact a large number of customers
Helps you optimize customer engagement and service strategies by revealing trends, opportunities, potential issues, and the root cause of customer perceptions so that you can take action quickly