Sorry to Bother You Again: Developer Recommendation Choice Architectures for Designing Effective Bots

Abstract

Software robots, or bots, are useful for automating a wide variety of programming and software development tasks. Despite the advantages of using bots throughout the software engineering process, research shows that developers often face challenges interacting with these systems. To improve automated developer recommendations from bots, this work introduces developer recommendation choice architectures. Choice architecture is a behavioral science concept that suggests the presentation of options impacts the decisions humans make. To evaluate the impact of framing recommendations for software engineers, we examine the impact of one choice architecture, actionability, for improving the design of bot recommendations. We present the results of a preliminary study evaluating this choice architecture in a bot and provide implications for integrating choice architecture into the design of future software engineering bots.