User recommendations for intelligent personal assistants

Adoption of intelligent personal assistants (IPA) is on the rise. Published studies on IPAs often focus on the analysis and critique of existing IPA features without understanding specific user needs that the technology aims to address. We present an exploratory study that gathered user recommendations for the design of their ideal IPA. The study relied on focus group and content analysis methods for data collection and analysis. Major themes in participants' recommendations for IPA design were identified and included feature improvements (e.g. speech recognition, input/output modalities, device feedback); customizability and increased control over IPA features and functions; transparency and understanding of IPA hardware, software, and data management processes; personification; compatibility with third-party platforms; accessibility; and aesthetics. Many of these recommendations are rooted in basic user experience design principles and have been previously discussed in the context of PDAs (personal digital assistants) and other technology. Addressing these recommendations would advance IPA technology and improve user experiences with it.