Vertext: An end-to-end ai powered conversation management system for multi-party chat platforms

Abstract

Online communication platforms like Slack and Microsoft teams have become increasingly crucial for a digitized workplace to improve business efficiency and growth. However, these chat platforms can overwhelm the users with unstructured long streams of back and forth discussions scattered in various places. Thus, discussions become challenging to follow, leading to an increased likelihood of missing valuable information. Moreover, with the unsatisfying keyword-based chat search, users spend a significant amount of time to read, digest, and recall information from the conversations at the cost of productivity. In this paper, we present Vertext, an end-to-end AI system that ingests user conversations and automatically extracts information such as announcements, task assignments, and conversation summary. Moreover, Vertext gives a unique search experience to the users by providing search results along with their context, with an improved performance enabled by semantic search. For the ease of user interaction, all the information is consolidated on a single dashboard provided by Vertext.