<p><strong>Knowledge workers are essential for sustainable organisational performance. However, knowledge workers face many challenges due to the increased demands of a globalised economy and rapid technological advancements. Moreover, research suggests that knowledge workers lack self-discipline and often engage in unproductive work habits, consequently impacting their productivity and wellbeing. Enterprise intelligent personal assistants (IPAs) are being promoted as having the potential to boost knowledge workers' productivity and wellbeing by regulating work behaviours. However, research on the role of IPAs in managing personal productivity and wellbeing remains underexplored within information systems (IS) research. Current studies are primarily confined to human-computer interaction (HCI) literature and predominantly centred on the IPAs' design and usability evaluation, neglecting to investigate their real-world impact and lived experiences. This thesis addresses this shortcoming by focusing on the underexplored aspect of human-AI coregulation of daily work-life practices. The study is based on a single unique case study of knowledge workers interacting with Microsoft Viva Insights (MVI), an enterprise IPA for productivity and wellbeing. The data for this study include twenty-six semi-structured interviews, four self-reflection journals, over sixty screenshots of interactions with MVI, twenty-two product videos and twenty-three webpages and documents on MVI. Using the concepts of coregulation and framing theory, this study explores both the role of enterprise IPAs and the nature of human-AI interaction in regulating daily work-life practices. I explain how the IPA, as a coregulator, used rationalisation, moralisation and normalisation strategies to participate in the regulation of work-life practices. In addition, I present a frame interaction model of coregulatory human-AI interaction to show and explain that for human-AI coregulation of work to be successful, there is a need for frame alignment between humans' cognitive frames and the frames embedded in IPAs. The significance of this study is that it adopts a novel approach by combining two bodies of framing literature and extending the application of the framing theory to explain the nature of human-AI interaction in the context of agentic IS designed for behaviour regulation and modification in work settings. As far as I am aware, this study is the first to merge two bodies of framing literature to conceptualise human-AI interaction as a form of frame interaction within the IS field. The study offers valuable empirical insights for IS researchers, AI designers and organisations interested in effectively integrating AI technologies in the workplace.</strong></p>