Review of visual programming environments for conversational interfaces

Abstract

Conversational interfaces are a rapidly growing trend among organizations and consumers around the word. It is our belief that the process of production of conversational interfaces is more efficient when developed visually. From Virtual Assistants, to chatbots, everything interacts with users through conversation, instead of relying on traditional Graphical User Interfaces. Our goal is to develop a programming language, and a corresponding tool for developers to visually design the flow of conversations for any application in the field of dialogue management, while being able to perform various operations such as mathematical computations, programming functions (logics, conditions, loops, etc.), database queries, and external API calls. The proposed tool could be used to develop applications for existing VA and chatbot platforms such as Amazon Alexa or Facebook Messenger, or novel applications for custom systems interacting with users through conversation. In this paper, we propose an architecture for such a system and review existing implementations for each of its components.