When 80% of your organisation's code is suddenly written by AI, the fundamental nature of software engineering changes. At 1KOMMA5°, our rollout of AI coding assistants to everyone fundamentally altered our culture and ways of working. We will take you on the journey of transforming our engineering organisation in only a handful of months, sharing our hard-won lessons with you.

In this talk, we will share what it took to drive the adoption of AI-coding across our organisation, including how and why we intentionally involved non-engineers like designers and PMs. We will explain how we established a culture around sharing best practices and professional use, highlighting the critical role of communities of practice playing in bridging the gap between management mandates and grassroots engineering adoption.

Beyond the cultural shift, we will detail why we intentionally restricted the budget during our coding assistant proof-of-concept to build developer empathy for LLM costs and model selection impact. We will also break down the exact data we collected, which data proved to be the most valuable, and why this transformation is forcing us to revisit traditional engineering health metrics like DORA. Finally, we are going to share the still-unresolved issues of introducing AI coding organisation-wide, and how we plan to resolve them.