Coding scales – Thinking remains limited

AI coding is not just a matter of tools – it is a strategic leadership task
Those who establish a structured, company-specific approach early on will gain sustainable competitive advantages – while those who experiment in a piecemeal fashion or simply copy others risk wasting time, causing organizational friction, and creating inefficiency.

Generative and agent-based AI are fundamentally transforming software development
What begins today as AI coding and vibe coding is evolving into a new model of digital innovation: AI-centric software development. Large language models and development agents can already accelerate or partially automate many software development tasks – from writing code and testing to documentation and refactoring. This drastically reduces implementation costs and time. We see the results in technology and tool development, in software and innovation processes, in the job market, and in the sometimes comparatively low valuations of software stocks.

From coding to AI-centric innovation: A structural shift in software development
We are developing and using new software development tools, but there is much more at stake – the structure of digital innovation is changing. Since implementation scales better through automation, it is no longer a central bottleneck. The focus is shifting from implementation to design: from coding to problem understanding, system design, architecture, and the orchestration of intelligent product and process development. Software development is thus evolving from an implementation-driven discipline into an AI-centered innovation system. Within this system, new bottlenecks, new roles, new processes, and new governance requirements are emerging.

AI-centered innovation: Hypotheses, open questions, and an invitation to dialogue
This discussion paper does not present established scientific findings, but rather questions and hypotheses. It reflects our current observations and thoughts as inspiration for others and serves as an invitation to dialogue. The field is evolving so rapidly that it is not feasible to outline a detailed vision for AI-centered innovation and software development over the next few years and then systematically implement it. On the other hand, one cannot ignore the new possibilities and opportunities without falling behind the competition. Companies should not view generative and agent-based AI merely as productivity tools for coding, but should instead align their innovation and development organizations early on with an AI-centered model of software development – with a stronger focus on problem understanding, system design, and governance – while learning and experimenting iteratively to quickly capitalize on opportunities without rigid roadmaps.