By now, everyone is familiar with the idea of using AI to code, from
helpful autocomplete through fully-autonomous vibe-coding. But software
development is more than just coding: how can we use LLMs to assist with
testing, deployment, and operations? What about with feature design, or
writing user stories? And where do Agile practices fit into AI-assisted
development: which practices are still needed, which require
modification, and which are no longer pulling their weight? This talk is
a survey of points in the software lifecycle where you can apply LLMs in
an Agile way right now, and an in-depth discussion of three specific
uses, including security review.