At the Software Excellence Network summit, Transform or Face Disruption, Prof. Dr. Wolfgang Pree, Professor at Universität Salzburg and CEO of AI², laid out a transformative vision for the future of enterprise IT. His core message was clear: the traditional paradigm of software development is giving way to a new era—one where AI models, not engineers, increasingly define how software is created, deployed, and evolved.
This shift is not theoretical. It’s already underway, and CIOs will need to rethink everything from architecture to team composition to remain competitive.
Pree highlighted a key distinction that is too often overlooked: while the underlying architectures of large language models (LLMs) may be well-understood in isolation, their emergent behavior cannot be easily predicted or controlled. This places them in the category of complex adaptive systems—more akin to biological organisms than deterministic machines.
For CIOs, this means moving away from assumptions of linear scalability and towards strategies designed for systems that evolve in unexpected ways. Traditional control structures and risk models may not apply.
The masterclass addressed a fundamental transformation already visible in enterprise environments: the rise of natural language interfaces that allow users to specify functionality without writing a single line of code. Several LLMs and AI tools are now capable of generating entire applications from textual instructions.
In educational settings, Pree shared examples of semester-long student projects being replicated by generative AI systems in just hours. This trend challenges the longstanding assumption that only trained engineers can build enterprise-grade software. It also suggests that software development could become far more accessible across the organization—empowering business units, operations teams, and even non-technical staff to automate their workflows.
Rather than enforcing strict top-down controls, Professor Pree advocated for enabling experimentation through structured sandbox environments. These should be time-boxed, governed by clearly defined goals, and designed to encourage trial and error. This approach allows organizations to test ideas rapidly while containing risk.
Startups have already demonstrated the effectiveness of this model, but Pree noted that large enterprises often struggle to replicate it due to cultural and bureaucratic inertia. CIOs who want to capture AI’s full potential will need to implement internal policies that prioritize speed, accountability, and adaptability over traditional project governance.
A major implication for CIOs is the evolving skill set required across IT teams. As traditional programming gives way to prompt engineering, model curation, and data-driven design, technical teams will need fewer syntax specialists and more professionals fluent in orchestrating and optimizing AI systems.
While conventional coding will still matter for some infrastructure-level work, the emphasis will shift toward managing data quality, monitoring system behavior, and adapting AI components within complex environments. This also means investing in AI literacy across the business—not just within IT.
Prof. Pree emphasized the unpredictability of AI systems, not only because of their complexity but also due to the rapid pace of change in model development. Capabilities that were considered impossible just a few years ago—such as logical reasoning or multi-step problem solving—have emerged seemingly overnight through unexpected model behaviors.
This underscores the need for CIOs to build AI governance frameworks that allow for fast iteration, failure, and course correction. It also means being ready to integrate new models quickly, adjust user interfaces or workflows, and constantly evaluate the evolving performance of AI components.
As AI systems become central to software production, there is a growing argument that computer science itself may evolve into a new kind of discipline—less focused on writing code and more concerned with teaching and guiding intelligent systems. If so, the organizational functions around software development will shift from technical implementation to behavioral oversight and system education.
While this is a long-term shift, the short-term implications are already visible. CIOs must begin adapting now by rethinking how they structure teams, invest in tools, and assess IT performance.
Prof. Dr. Pree’s masterclass ultimately urged CIOs to abandon the idea that software engineering is a static discipline. Instead, IT leaders should prepare for a future where the most critical role is not building systems—but shaping the environments in which intelligent systems learn, adapt, and thrive.
Note: While I've attempted to stay faithful to Prof. Dr. Wolfgang Pree's message, this is of course just my interpretation of it. I recommend watching the professor's full presentation to understand the context and depth of his remarks. If you have any questions or concerns, do not hesitate to reach out to me.