Question and Answering (QA) systems are referred as virtual assistants and are envisioned to be the next generation call center. However the accuracy of such QA systems is not as desirable and needs significant enhancement. Understanding the intent of the query is a significant contributor to an efficient system which has not been often analyzed. The current study intends to develop a QA system which can understand the query intent by using NLP based classification along with a novel scoring mechanism to extract the related information. Initially an insurance domain based simple frequently asked QA system was developed, which was then operationally enhanced into a system that can answer any type of insurance related query. The system was tested with 200 different queries and the output was cross validated by 3 experts from insurance field