Today’s service desk team would be receiving about hundreds of queries via messages, emails and calls per day. Thus there is a great need for more and more human manpower for routing of these queries towards concern department, which are supposed to handle those particular queries. The processes of routing queries require manpower, which is not that accurate because of various reasons. Thus, this problem requires a need for automation of receiving queries at the service desk and then routing these queries to appropriate department.

This paper explains the automation of the queries using machine learning algorithms. In this paper, emphasis is made to identify the best algorithm, which can be used for classification of queries. The various algorithms used are Multi-class Neural Network, Decision Tree, Multi-Class Logistic Regression Techniques, etc. From the analysis we obtain best model accuracy in Multi-Class Neural Network Method,