The support team at Zuora needs to actively involve human agents for the resolution of support tickets. If the support ticket is related to a problem with Salesforce.com process, the support agent needs to request Salesforce.com org access to the customer. Requesting access and getting it requires a lot of back and forth thus adding to the ticket resolution time. If the time required to get the Salesforce org access is reduced, it will lead to a significant reduction in the overall problem resolution time.
In this project, we aim to use Machine Learning to solve this problem by using the concept of Correlated Topic Models, specifically the Pachinko Allocation Model.
Technologies -
- Python 3.x
- Java v1.7
- MALLET/Spacy (NLP)
- Tensorflow
- Neo4J (Graph DB)