Understanding householder behaviours through the lens of a human assistant metaphor to inform the design of the in-home IPA virtual agent

This thesis explores the role of the intelligent personal assistant (IPA) device (e.g., Amazon Echo Dot) within the research setting of the home and outlines human-centred design recommendations towards the advancement of domesticated virtual agents (e.g., Alexa). 

Chapter 1 introduces the thesis and outlines how the subsequent chapters will be structured. Chapter 2 identifies the challenges presented within IPA research literature when studying their use within the home: (1) the lack of longitudinal enquiries documenting how in-home IPA use evolves; (2) the absence of considering meaning embedded within the home, when studying the use and usefulness of an in-home IPA virtual agent; and (3) the focus on interaction-level information, with the user experiences that these IPA virtual agents can provide being largely overlooked. This thesis investigated the user interactions and experiences with a gifted in-home IPA device through the research narrative and metaphor of employing a human assistant for the home. This was undertaken by applying a roleplay approach throughout the four research case studies, which depicts the key touch points of the employment lifecycle of a human assistant (Chapter 3). Research methods appropriate for the issues identified in the literature are outlined in Chapter 4. An exploratory study (framed by the research narrative as a household audit) captured the everyday tasks and activities that form householders’ domestic routines. Similarities of the routine structure were apparent across all households, yet how these tasks and activities were approached and undertaken differed when viewed through a temporal perspective (Chapter 5). Based on the outcomes of Chapter 5, each research participant was gifted a suitable in-home IPA device (Chapter 6). The subsequent study (Chapter 7) was framed as a performance review to document the lived experiences of owning and using an in-home IPA virtual agent. This identified that IPA virtual agents could, to some degree, ease domestic responsibilities for a householder. Chapter 8 was framed as an educational training session, established to broaden the research participants’ understanding of the capabilities of their in-home IPA virtual agent; it was evident that participants were willing to learn new commands for functionality that would support their domestic responsibilities. Chapter 9 concluded the research enquiry and was framed as the ‘cessation of employment’ - this allowed the research participants to express their account of living with an in-home IPA virtual agent for 30 weeks. Outcomes were that all research participants, but one would recommend an in-home IPA virtual agent to a prospective owner, describing a variety of positive or neutral overall experiences. Based on the study findings (Chapters 5-9), this thesis discusses (Chapter 10): (1) householder responsibilities that were delegated to an in-home IPA virtual agent; (2) motivation to use derived from the capabilities of an in-home IPA virtual agent; (3) how time-saving capabilities of an in-home IPA virtual agent can support householders in completing their domestic routines; and (4) whether context awareness is necessary for an in-home IPA virtual agent. Human-centred design recommendations for enhanced in-home IPA virtual agents are outlined in Chapter 11 to support the design-driven domestication of these virtual agents. Chapter 12 concludes the thesis, including the extent to which the thesis aims and research questions were addressed, identification of future work and description of the study limitations.