AI makes coding faster than ever. But speed was never the real problem.
In most projects, code still becomes the source of truth. Requirements drift, documentation rots, and developers spend more time fixing misunderstandings than building solutions. Faster typing does not fix that.
This talk presents a different approach: the AI Unified Process (AIUP) https://aiup.dev. Instead of starting with code, we start with clear, structured requirements. These requirements become the central artifact. From them, AI generates diagrams, models, code, and tests. When requirements change, the system is updated in a controlled and consistent way.
You will see how to write requirements that AI can reliably work with, and how to keep business intent, architecture, and implementation in sync over time. The talk is based on real-world projects and shows concrete workflows, not theory.
AI gives software engineers the chance to do what they should have been doing all along: design systems, understand the domain, and build the right solution. Not just more code.
