The integration of generative AI into platform engineering offers significant opportunities to optimize and automate workflows, especially in regulated software development contexts. Unfortunately, embedding a continuous generating cycle can result in hours of sometimes excessive processing and production that does not produce value. In this talk, we will look at how to take a zero-trust approach to ensure security and compliance, without affecting processes already in place, and preventing massive production of unnecessary items. Through practical strategies, we will propose how generative AI can be incorporated into platform engineering in regulated scenarios by letting the user approve the deliverable without burdening it with significantly more work than normal, avoiding the generation of broken code for review. Preferring instead to focus on efficiently addressing priority critical reports and filling key gaps in test coverage.