Virtual Assistant for Regulation-Dense Organizations

Abstract

Virtual assistants have become a large part of our daily lives. There are services such as Amazon's Alexa, the Google Home, Apple's Siri and Microsoft's Cortana. All of these provide information simply through the user's voice commands. This is useful technology and makes our lives easier. However, existing virtual assistants have provided generic information, not domain-specific information. For examples, if you ask Siri a very specific question about events in your company, Siri will most likely not able to return a correct answer because it relies on a result of Google search and in many cases, Google does not know domain-specific information. To address this issue, we develop Slemp, a virtual assistant for regulation-dense organizations that possess a large quantity of specific and unique rules or guidelines. Using Slemp, users can easily feed domain-specific information and use their own virtual assistant to provide specific answers to users in their organizations. In this paper, we demonstrate the design and development of Slemp for our institute, Virginia Military Institute, and our work cannot be easily adopted by other similar organizations.