You can get a working app in a few minutes now. An AI assistant, a quick prototype, a generated codebase. For a first version, that works well. What happens next is where it gets interesting. There's a long distance between something that runs on your machine and something a company actually uses in production, and that distance is where most projects have always run into trouble. Writing code was rarely the real obstacle. The problems showed up somewhere else. Who is allowed to do what. How the process actually runs, not how someone thinks it runs. Which business rules apply, and which exceptions nobody mentioned until the system hit them. This is seldom clear at the start, and even when it is written down, people read it differently. As the cost of producing code falls, the hard part becomes what it always quietly was: understanding the domain, modeling the system, and designing how it should behave in the real world. In this session I'll look at two ways teams speed up building today, AI-assisted development and low-code platforms, less as competitors and more as tools that suit different jobs. We'll spend most of the time on the part that decides whether software gets used at all: the work around the code, from analysis and requirements to testing, change, and adoption. The engineers who matter won't be the ones who type fastest. They'll be the ones who understand how a system needs to work in practice. That was true before AI, and it's more true now.