Cybernetic Delivery Method™ adding value when tackling legacy software challenges

The challenge: A legacy system no one could fully explain

When the Zühlke team joined the client’s project, it faced a system that worked, but no one knew exactly how. Change was risky. Maintenance was expensive. And despite its importance, the system could not keep up with the pace of business.

Millions of lines of legacy code, no reliable documentation, and business logic scattered through years of exception-based rules made the client’s digital landscape a black box. It was fragmented, complex, and held together by assumptions no one could trace.

The goal: A modern, AI powered solution to enhance clarity, speed and scalability

Support the client in fully rebuilding & modernizing the legacy technology stack to bring clarity, speed, and scalability back to the heart of its operations.

With our long-standing business relationship, the client chose Zühlke based on our track record in successfully reconstructing intricate systems, paired with the expertise to ensure continuity, stability, operational efficiency and scalability.

The solution and tools

Faced with this intricate challenge, the Zühlke team decided to leverage AI to augment our methods of working, addressing complexities, accelerating processes, and laying the groundwork for future innovation and sustained competitiveness.

We implemented our newly developed Zühlke Cybernetic Delivery Method™ (CDM) which defines Zühke’s approach towards AI augmented delivery of digital products and services. Cybernetics, the science of systems, control, and communication between people and machines, provides foundational principles that can directly support responsible AI use.

At the heart of CDM is the idea of combining human expertise with advanced AI assistance, enabling faster, more effective delivery of digital products and services with greater agility and impact.  

With the CDM Patterns, we have developed reusable solutions for effectively integrating AI support into digital product and service delivery, improving productivity across disciplines. They help teams move from task-level automation to transformative workflow improvements, ensuring that AI-driven acceleration delivers real business impact. 

Finally, one of the key aspects of CDM are cross-functional communities that drive continuous learning by capturing best practices and real-world insights, helping teams adapt quickly to new AI technologies. This ensures faster innovation, responsible adoption, and sustained competitive advantage.  

As anticipated when the team started applying CDM, they have found these steps to actually improve productivity and reported work as being ‘more joyful’: AI is well-supervised, eliminating most routine work so the teams can focus on innovating much faster. Humans decide what is the right thing to do and AI does things right.

Here’s how the team is working with AI today:

• Business Analysts & Architects use ChatGPT to reverse-engineer undocumented legacy code and complex data structures. Business Analysts are enabled with tools like SQL-Query GPT to help them question the database in the natural language or Story-GPT to generate new user stories with the standardized formatting.
• Developers benefit from tools like GitHub Copilot, ChatGPT, and custom GPTs enriched with the team’s coding guidelines. These tools, supplemented with team context speed up development, ensure consistency, and explore unfamiliar parts of the system. Code reviews are supported by the Review Bot, catching style and quality issues earlier in the process.
• Quality Assurance (QA) plays a crucial role in the project, with AI actively supporting quality assurance throughout development.  CDM patterns combined with the power of tools like ChatGPT and GitHub Copilot are used to generate additional test cases and act as sparring partners to identify potential gaps in coverage – helping the team catch edge cases early and maintain high confidence in their deliverables. 

The results

All in all, leveraging CDM as an underlying methodology for augmenting the software delivery lifecycle with AI, the team now:
 • Works smarter, from planning to implementation
 • Codes faster, with improved quality and fewer manual errors
 • Automates the repeatable tasks – freeing up time for high-impact thinking

Outcome: team reported ~30% efficiency gains when developing this modern, maintainable platform with more confidence and clarity.

Looking forward, the new system now serves as a scalable foundation for future innovations at the client’s organisation, including predictive analytics and self-optimizing workflows.