Creating Routines for IoT Ecosystems through Conversation with Smart Speakers

Nowadays, end users can create routines for Amazon Echo and Google Nest devices using a companion app (Amazon Alexa and Google Home, respectively) running on smartphones. Our work explores the possibility of transferring this End-User Development activity directly to the smart speakers, with and without a touchscreen. To this aim, we designed and developed two Amazon Skills (one for Amazon Echo Show and the other for Amazon Echo Dot) and two Google Actions (one for Google Nest Hub and the other for Google Home Speaker). Then, we carried out two controlled experiments, involving 40 participants, to compare routine creation through multi-modal interaction (based on vision, speech, and touch) with routine creation through speech-only interaction. Driven by our research questions, we found that for routine creation the multi-modal interaction is preferred to the speech-only one and the perceived quality of interaction seems to depend on the brand of the smart speaker.