Success of an Agent-Assisted System that Reduces Email Overload

Andrew Faulring, Carnegie Mellon University
Brad Myers, Carnegie Mellon University
Aaron Steinfeld, Carnegie Mellon University

Can be found at

http://repository.cmu.edu/hcii/264

Abstract or Description

RADAR is a large multi-agent system with a mixed-initiative user interface designed to help office workers cope with email overload. RADAR agents observe experts performing tasks and then assist other users who are performing similar tasks. The Email Classifier learns to identify tasks contained within emails and then inspects new emails for similar tasks, which are presented in a novel task-management user interface. The Multi-task Coordination Assistant learns a model of the order in which experts perform tasks and then presents subsequent users with a suggested schedule for performing their tasks. A large evaluation of RADAR demonstrated that novice users confronted with an email overload test performed significantly better (a 37% better overall score with a factor of four fewer errors) when assisted by both agents. Additionally, in a post-test survey users perceived the test to be significantly more difficult when they did not receive assistance from both agents, indicating a preference for the AI-based assistance. We also observed a wide variation among users in the amount of agent advice that they followed.