AI-Generated Code in Practice

A test automation example

This page presents a focused example application whose code was generated using AI.
The application itself does not use AI at runtime.

AI was used purely as a development tool during implementation.

If you’d like to discuss how AI-generated code could be analyzed in your own environment, happy to talk.

What was built

What this application does

- Web application test automation
- Recording tests based on user interactions
- Replaying tests later as part of integration or regression testing
- Playwright for test execution
- OCR and text recognition for UI observation when DOM-based detection is insufficient

This is not a commercial product, but a deliberately scoped technical example.

How AI was used — and how it was not

AI was used to:
- generate application code
- accelerate repetitive development work

AI was not used to:
- create tests automatically
- make runtime decisions
- operate the system in production

Architecture, boundaries, and behavior were defined and validated by humans.

Why AI-generated code still needs analysis

AI can generate large amounts of code very quickly.
That alone does not guarantee that the result is:
- safe
- maintainable
- scalable
- long-term viable

This example illustrates why AI-generated code always requires proper analysis, validation, and technical judgement before production use.