nowledge interface systems (KIS) enable a dialogue between human and machines by utilizing an underlying knowledge-based system. The design and the effects of KIS have been a focus of researchers' interest for decades. Yet the existing knowledge of KIS is scattered, and researchers as well as practitioners face the danger of re-inventing KIS for a specific purpose or, worse, repeating mistakes of the past. This conceptual paper provides an overview of KIS capabilities and outcomes of their usage from the past to the present and proposes directions for future research. Our analysis shows that, in general, there is evolution of work on KIS over time, rather than revolution. This research will enable researchers to identify their contribution more clearly over and above what has been done before.