Will Generative AI Fill the Automation Gap in Software Architecting?

Researchers are aware that software architects lack effective automation to support much of their work. Generative AI (GenAI) is sparking research interest regarding its potential role in filling this gap, inspired by promising applications of GenAI to other software engineering activities. In this paper, we aim to reflect and sharpen this conversation from the vague “how can GenAI be applied to architecture” to “which architecture activities are most amenable to application of GenAI.” We stress the importance of considering contributions in the context of workflows and reflect on the alignment (or lack thereof) of GenAI with the nature of common architecture tasks through the discussion of five common architecture activities. We offer guiding criteria to assist architecture researchers in focusing on activities that are both amenable to automation and likely to obtain significant utility from GenAI.