An architecture roadmap for introducing an AI-DLC.
For a large Swiss company, I designed the roadmap for an AI-Driven Development Lifecycle (AI-DLC) as an architect: a development process in which AI carries out steps from requirements to operations and people review the results.
Situation
Introducing AI more widely in software development.
AI tools were already in use, with individual employees gaining experience. Wider adoption required agreed practices and technical provisions, particularly for reviewing results and integrating AI into the development process.
I worked with leadership and the people responsible for architecture, software development and security. To preserve confidentiality, this account omits the client name and internal metrics.
My role in the project
Test AI together
I demonstrated what AI can do through typical work situations. Leadership and technical owners used this to discuss where to introduce AI and which tasks would remain with employees.
Define tasks and responsibilities
I helped the participants compare how independently AI should carry out tasks and which checks people should perform. We examined the differences in costs, responsibilities and risks across the possible operating approaches.
Plan the architecture
I planned the architecture for introduction and subsequent expansion, accounting for AI potentially performing further tasks independently and the technical changes this would require.
Define safeguards
I clarified security requirements for the intended use. Together with the responsible people, I documented which measures were needed before introduction and which were planned for later.
Outputs for project planning
The roadmap included the approach to software development with AI, the planned architecture and the requirements for further adoption. Leadership and technical owners could use them to plan implementation and assess the implications for costs, security and operations.