Research-in-the-loop is an approach to integrating UX research into AI-assisted software development. It involves specifying research requirements alongside product features and using coding agents to build instrumentation and evaluate releases. This approach allows researchers to inform development decisions, investigate problems, and discover new user needs. Research-in-the-loop creates new types of work for researchers, such as writing research requirements, directing coding agents, and building tools. However, many questions remain open, including where different forms of research fit, how to monitor and evaluate system-level effects, and how to design instrumentation to derive sound conclusions. https://www.nngroup.com/articles/research-in-the-loop/


