AI Disclosure
How I use AI in software development
I use AI coding assistants to build faster. The responsibility for what ships stays with me. This page explains where the line is.
What I own
- Requirements and scope
- Architecture and design decisions
- Integrations with systems such as Odoo and Microsoft 365
- Code review, testing and validation
- Deployment, backups and ongoing operation
Where AI helps
- Writing code faster to my specifications, using AI assistants such as Claude (Anthropic), ChatGPT (OpenAI) and Gemini (Google)
- Building parts of web applications with Lovable's AI editor, under my direction
- Choosing the tool per task; the same review and testing rules apply to all of them
How the work is checked
- AI-generated code is treated as a draft until it passes my review and the project's automated tests.
- Each case study states the project's real status: in production, pilot, in development, prototype or concept.
- Figures such as timelines or test counts are only published when they can be verified.