Customized Automated Email Response Bot using Machine Learning and Robotic Process Automation Robonomics AI India Private Limited, L-1804, Purva Highlands, Survey #19, Mallasandra Village, Holiday Village Road, Kanakapura Road, Bengaluru- 560062, India. info@robonomics.ai Maharsh Patel Department of Information Technology K J Somaiya Institute Of Engineering and Information Technology, Sion, Mumbai, University Of Mumbai, maharsh.p@somaiya.edu Raunaq Porwal Department of Information Technology K J Somaiya Institute Of Engineering and Information Technology, Sion, Mumbai, University Of Mumbai, raunaq.p@somaiya.edu Aastha Shukla Department of Information Technology K J Somaiya Institute Of Engineering and Information Technology, Sion, Mumbai, University Of Mumbai, aastha.s@somaiya.edu Radhika Kotecha Department of Information Technology K J Somaiya Institute Of Engineering and Information Technology, Sion, Mumbai, University Of Mumbai, radhika.kotecha@somaiya.edu Abstract— Email is one of the most reliable, dependable, authentic, genuine and frequently used modes of communication in corporate world. It is a common practice in corporate culture to send an automatic response to new incoming messages but it is very generic and limited in extent of its utility. A system can be designed to substitute the tedious task of replying to thousands of mails manually. This work proposes a system which in broader context automates the task of responding to a client. It aims to mitigate the laborious task of perusing through email manually by customizing the response as per the query of the user and fabricating a reply accordingly. The proposed system proves the effectiveness of the approach in its implementation. Keywords — Email, Robotic Process Automation, Associative Learning Algorithm, Support Vector Machine, Text Rank Algorithm, Email Classification. I. INTRODUCTION One of the most ubiquitous and widely used modes of communication in professional as well as personal aspects is Email. It is always exciting when one seeks someone or some enterprise out via email and they respond instantly in no time. [1] In today’s society, especially as technology grows, people are looking for immediate response and satisfaction. Not only does it provide with a boost in efficiency but also it can be a great and a very useful business tool if used properly. For example, if you cannot respond within 24-48 hours, automated emails [1] are great ways to inform the email recipients when you will have a delay in your response due to a project, vacation, and so on. Email is amongst the most robust and secures means of online communication and has become communication tool of paramount importance for most enterprises as well as professionals. [2] With the usage of email flourishing at a rapid pace, management and tracking of emails becomes an intimidating task. A median user spends a healthy amount of time in reading, perusing and responding to such incoming emails. Furthermore, majority of the emails follow a fixed structure and pattern in terms of content and require simple replies in most cases. [2] E.g., a recent study projected an email reply system for an organization’s Frequently Asked Questions (FAQ). In FAQ’s, a lot of the questions are repeated by multiple users, thus making them redundant. By making use of proper classification, correct and dynamic responses can be formulated to answer the incoming user queries [1]. To enhance user’s experience by automating email replies, with appropriate classification and timely scheduling, this paper presents a novel solution [1] for formulating automated reply to emails on a priority basis after proper classification. This solution will possess the capacity of accurate email reply prediction employing both, Supervised [3, 4] as well as Unsupervised Learning modules [1, 2]. The proposed solution will be able to produce correct, highly formatted and time stamped emails. It will also act as a catalyst in development of a system for email reply system with primary feature of 2nd International Conference on Advances in Science & Technology (ICAST-2019) K. J. Somaiya Institute of Engineering & Information Technology, University of Mumbai, Maharashtra, India http://ssrn.com/link/2019-ICAST.html Electronic copy available at: https://ssrn.com/abstract=3370225