Vincent Slevin
 Design Conversational Services

PhD
TechnologyHuman-Computer InteractionSoundscapeConversational Interaction DesignService DesignCyberneticsConversation AnalysisUsability FactorsExperience FactorsInteraction Design

    Summary

    Conversational User Interfaces (CUIs) are used in a growing number of different contexts from web based chatbots through to entertainment devices in homes, in cars, in museums, and in shops. Organisations are adopting CUIs to meet the growing demand for end-to-end self-service experiences that are convenient, frictionless, and accessible 24/7 across all digital channels in a seamless way. However the lived experience of these CUIs often fails to meet expectations (Reeves 2019) as the systems they operate on rely on reductionist models based more on past data than on contextual cues.

    This project explores the impact that design research methods can have in the emergent field of CUI design. As a piece of research-through-practice (Frayling 1993) it has shaped the design practice of a conversational AI design studio, through the act of researching and prototyping conversational services for everyday use.

    The question of how to humanise conversational AI technology for the purposes of everyday services is an ill-defined, ill-structured, messy real-world issue that is of both practical and scholarly importance; and one that is appropriate for methods of design research. Through a conscious bid to re-imagine Alexa as an everyday computing interface for online food ordering, this research demonstrates the contribution that different design research methods can make in innovating AI-based services.

    Additional info

    This project took a systemic approach to the field of CUI design involving users, merchants, technology infrastructure suppliers, designers, investors and developers in a rolling state of the art review. The design research methods used were: action research, human centred design, usability factors, and conversation analysis. These methods have been applied across three successive phases of research-through-practice.

    Prototype iterations have been shaped and critiqued across each phase with increasing fidelity and scale. This has generated a set of design artefacts that reveal novel considerations for CUI designers, and demonstrate the impact that design research methods can have in CUI design practice.