Automating customer complaints classification in German
Who is this presentation for?
- Data science managers, big data use case managers, and data scientists
When a customer complains about an issue, the length of time it takes to answer the query and provide a solution depends on getting to the right agent handling the issue. For example, an agent might be handling only mobile contracts, and if the agent receives a complaint about an internet contract, then the request has to be rerouted to the right agent. Depending on the time and availability of the right agent, the customer complaint might take a few days for resolution. The agent also has to read the German text and classify the customer complaint text to a topic—going manually through a tree of a 56 predefined topics. A manual classification costs time and money, as the agent will bill for the hourly classification, costing not only thousands of euro but also a delay in customer response. Bad service often leads to customer churn.
Adithya Hrushikesh explains how, in order to address the time, churn, and classification costs issue, the company uses natural language processing techniques on Twitter and Facebook in German to automatically classify the customer complaint text in the data warehouse to one of the predefined topics.
- A basic understanding of data science
What you'll learn
- Learn how to build a data product and deploy the model in production
- Understand ensemble models, design thinking in data science projects, building a business case, and GDPR
Adithya Hrushikesh is an operational intelligence lead for data science at Vodafone, where he leads a team of data scientists and data engineers to build data products.
For conference registration information and customer service
For more information on community discounts and trade opportunities with O’Reilly conferences
For information on exhibiting or sponsoring a conference
For media/analyst press inquires