We report work in progress that examines the neurophysiological and perceptual effects upon users of a Generative AI tool developed to serve as a general assistant within an enterprise, helping to perform various administrative tasks. Participants interacted with the AI tool in three ways: chatbot, chatbot plus summary, or chatbot plus form, with an additional manual control task. We compared neurophysiological and perceptual responses across tasks and, more broadly, across AI interaction modalities. We performed time-frequency decomposition to derive power estimates in the delta, theta, alpha, beta and gamma frequencies for all tasks and interaction modalities, producing t-maps to illustrate the differences in brain activity. Our findings illustrate that interacting with AI requires varying degrees of cognitive effort compared to manually performing tasks. However, the dynamics of this brain activity are nuanced and vary significantly across interaction modalities, as do perceptions.