Conversational agents are coming to life on all kinds of devices that include any possible combination of display sizes, microphone capabilities, speaker quality, etc. For multimodal interactions designed to operate on a wide range of surfaces, the UX should account for all of these factors to achieve a consistent sound-and-feel and look-and-feel. This talk will delve into what to consider when defining an interaction model that will allow these experiences to adapt to any conceivable surface.