What Can You Do For Me? The Discoverability of Intelligent Assistant Skills

Manveer Kalirai, Anastasia Kuzminykh

Abstract: A wide discrepancy exists between the range of available Intelligent Personal Assistant (IPA) skills and the range of skills users engage with regularly. One reason for this is the issue of low skill discoverability. From the literature, we understand that there are factors that can potentially enhance discoverability, such as context of use. The literature also signals that current discoverability strategies, which leverage such factors, are being challenged by users’ privacy concerns, and rapid skill growth. Mindful of these challenges, we explore the ways users naturally group IPA skills as a possible springboard for discoverability. Our preliminary findings suggest that grouping skills based on users’ functionality needs in different areas of life may be a viable direction toward advancing discoverability.