For years, we have been cutting back on good engineering practices to deliver on time. Now that we can delegate much of the implementation to AI, we have an opportunity to review, and an obligation to improve, our processes from end to end. Working with AI in real projects means rethinking many things: what a project needs to work with agents, what agents need to adapt to the project, how to ensure technical decisions are respected, and how to validate that what is delivered matches what was expected.

This talk shares practices being applied and approaches still being refined. It covers agentic development, Spec-Driven Development, hard and soft guardrails, testing and verification evidence, and the evolving concept of software product definition. Real examples show the advantages and drawbacks of AI at each stage of the process and what still depends entirely on engineers. AI is changing software engineering, and these practices illustrate how some of us are shaping the way we will work in the years ahead.