Uncovering the to-dos hidden in your in-box

Abstract

In this paper we present SCOUT, an application that examines the machine-generated messages within the in-box of an e-mail application, extracts from these messages information regarding the tasks the recipient is asked to perform, and displays these messages in a graphical interface where they are grouped by context. The tool is intended for business managers who receive daily a large number of machine-generated messages that require some action be taken. SCOUT uses the IBM Unstructured Information Management Architecture (UIMA) framework to apply rule-based reasoning for identification of tasks, and it uses contextual data to customize the presentation of task information to the user. SCOUT's open, extensible architecture allows the use of alternate inference models (such as machine learning algorithms) as well as the integration of additional context sources and client interfaces. SCOUT was well received by the participants in a small evaluation study.