AI-Driven Software Engineering Transformation at Getaway Group
An exploration of the profound shift in software development processes and organizational structures. Featuring Bastian Buch, CPTO of Getaway Group, who discusses the transition from legacy systems to AI-integrated workflows and the creation of an AI maturity model.
The New Era of Software Development
Software engineering is undergoing a fundamental change, shifting from traditional coding to a hybrid model where the focus is as much on building the 'machine that builds the product' as it is on the product itself. This shift is not merely a tooling change but a comprehensive organizational transformation.
From Legacy to AI-Integrated Workflows
At the Getaway Group, the journey began with a massive legacy load—PHP and FTP servers from 20 years ago. The transition didn't stop at cloud migration; it evolved into an AI-first approach. This includes the implementation of specialized AI skills and a 'Skill Marketplace' where internal developers create and share AI-driven automation tools.
One of the most impactful implementations is the 'Visor' system, a data analytics AI that allows users to query their entire data warehouse via natural language, effectively democratizing data access and reducing the reliance on manual SQL queries and Excel spreadsheets.
Redefining Roles and the AI Maturity Model
To combat the skill gap and prevent stagnation, the organization has completely overhauled its role descriptions for engineers and designers. Every role is now evaluated against a 10-step AI Maturity Model. This creates a transparent path for growth, providing employees with a structured development plan—curated by AI—to move from basic usage to advanced orchestration of AI agents.
Agile Evolution: Workstream-Based Organizations
Traditional, large-scale teams are becoming obsolete in the laelight of high-speed AI development. The organization is moving toward 'Workstream-based' structures: small, time-bound (max two weeks) cross-functional teams that focus on specific deliverables. This 'Project and Workstream Marketplace' allows for dynamic resource allocation and increases the speed of iteration.
Conclusion
The transition to AI-driven engineering is a journey of both technical adoption and psychological management. By combining high-risk appetite with clear guidelines and structured learning paths, companies can bridge the gap between legacy systems and the future of autonomous software development.
Key insights
-
The role of a software engineer is evolving into a dual responsibility: building the product and building the machine that builds the product.
Impact: This shifts the focus from manual coding to system orchestration, potentially increasing development velocity exponentially.
-
Traditional organizational structures and large teams are too slow and expensive for the current speed of AI-driven development.
Impact: A move toward small, time-bound workstreams allows for more agile responses to technological shifts.
-
An AI Maturity Model is necessary to provide a transparent growth path for employees, preventing fear and stagnation during rapid technological shifts.
Impact: Reduces workforce anxiety and ensures the company maintains a competitive technical edge.
-
AI Agents are moving beyond simple chat interfaces toward autonomous instances that can handle specific business flows, such as bug fixing end-to-end.
Impact: Could lead to the automation of 90% of routine maintenance and bug fixing, freeing humans for high-level architecture.
Action items
-
Implement a Skill Marketplace where developers can build and share custom AI 'skills' or plugins to avoid redundant effort and standardize automation.
Impact: Accelerates the adoption of AI tools across the organization and encourages a culture of internal innovation.
-
Develop a multi-step AI Maturity Model for technical roles to define what constitutes 'AI-competent' at different levels of proficiency.
Impact: Standardizes expectations and provides a clear roadmap for employees to upskill.
-
Transition from permanent large teams to small, time-bound (e.g., 2-week) cross-functional workstreams focused on a single deliverable.
Impact: Increases agility and reduces the overhead associated with traditional team meetings and reviews.
Quotes
“Die sind jetzt alle komplett neu gebaut und die sind alle, haben alle so twofold. Das eine ist, die Verantwortung, das Produkt zu bauen, und auf der anderen Seite die Maschine zu bauen, die das Produkt baut.”
“Ich glaube, ich bin mir sicher, dass ich mir in ein paar Monaten wieder unterhalten, will jammer schon einmal sehr spannend.”
“Wir müssen, ich glaube, die erste Option, wir müssen auf eine Sache setzen, so stark opinionated, aber echt versuchen, so weit wie möglich unabhängig zu bleiben.”